Content area
Artificial intelligence is emerging as a transformative force in addressing the multifaceted challenges of food safety, food quality, and food security. This review synthesizes advancements in AI-driven technologies, such as machine learning, deep learning, natural language processing, and computer vision, and their applications across the food supply chain, based on a comprehensive analysis of literature published from 1990 to 2024. AI enhances food safety through real-time contamination detection, predictive risk modeling, and compliance monitoring, reducing public health risks. It improves food quality by automating defect detection, optimizing shelf-life predictions, and ensuring consistency in taste, texture, and appearance. Furthermore, AI addresses food security by enabling resource-efficient agriculture, yield forecasting, and supply chain optimization to ensure the availability and accessibility of nutritious food resources. This review also highlights the integration of AI with advanced food processing techniques such as high-pressure processing, ultraviolet treatment, pulsed electric fields, cold plasma, and irradiation, which ensure microbial safety, extend shelf life, and enhance product quality. Additionally, the integration of AI with emerging technologies such as the Internet of Things, blockchain, and AI-powered sensors enables proactive risk management, predictive analytics, and automated quality control. By examining these innovations' potential to enhance transparency, efficiency, and decision-making within food systems, this review identifies current research gaps and proposes strategies to address barriers such as data limitations, model generalizability, and ethical concerns. These insights underscore the critical role of AI in advancing safer, higher-quality, and more secure food systems, guiding future research and fostering sustainable food systems that benefit public health and consumer trust.
Article highlights
AI improves food safety by detecting contamination risks and enhancing traceability across supply chains.
AI helps optimize food processing methods, reducing waste and improving efficiency in modern food systems.
Future AI innovations focus on scalable and cost-effective solutions to enhance global food security and sustainability.
Introduction
Ensuring food safety, food quality, and food security has become a complex challenge in the modern era due to the globalized food supply chain and evolving production methods. While these terms are sometimes used interchangeably, they address distinct but interconnected dimensions of food systems:
Food safety focuses on protecting consumers from foodborne illnesses and harmful contaminants, such as bacteria, viruses, pesticides, and heavy metals. It involves rigorous monitoring and management practices to minimize risks across the food supply chain.
Food quality ensures that products meet the desired standards for texture, taste, appearance, and nutritional value, catering to consumer expectations and maintaining consistency in production.
Food security pertains to ensuring that food is accessible, available, and safe for consumption worldwide, addressing issues of affordability, sustainability, and equitable distribution.
These three aspects, although distinct, are deeply interdependent. Unsafe food compromises quality and security, while inaccessible or insufficient food undermines security regardless of its safety or quality. These challenges are magnified by pressures from growing populations, climate change, and evolving consumer demands [1].
AI is emerging as a transformative solution to address these complexities, offering tailored applications for each dimension of food systems. By leveraging advanced technologies like machine learning (ML), natural language processing (NLP), computer vision (CV), and reinforcement learning (RL), AI enhances food safety through real-time detection and prevention of contamination risks. AI also supports food quality by enabling automated defect detection, optimizing processing parameters, and maintaining consistency across batches. Additionally, AI strengthens food security by improving resource management, yield prediction, and supply chain transparency.
Moreover, AI-driven optimization of advanced food processing techniques, such as high-pressure processing (HPP), ultraviolet (UV) treatment, pulsed electric fields (PEF), and cold plasma, ensures consistent safety and quality. Techniques like ultrasonic sterilization and irradiation can benefit from AI's ability to dynamically adjust key parameters, improving energy efficiency and reducing waste, thus enhancing the shelf life of food products while maintaining their safety and quality.
Figure 1 presents an overview of the key AI methodologies that are being used in food systems, along with the challenges and considerations that must be addressed for their successful implementation. The AI methodologies highlighted in this review include:
Machine learning (ML) – AI algorithms analyze large datasets to predict food safety risks, enabling early detection of contamination and spoilage.
Natural language Processing (NLP) – NLP helps monitor compliance by analyzing inspection reports, regulatory documents, and consumer feedback to detect food safety risks.
Computer vision (CV) – Image analysis tools are used to detect defects, monitor contamination, and assess product quality automatically.
Risk prediction – AI-driven predictive models allow for the anticipation of food safety hazards, facilitating proactive interventions.
Quality control – AI systems automate the detection of defects in food products, ensuring consistent quality standards.
Data analysis enhancement – AI accelerates data processing and analysis, improving decision-making for regulators and food producers.
Fig. 1 [Images not available. See PDF.]
Brief overview of AI methodologies and challenges in food systems
However, the successful implementation of AI in food systems is not without challenges. As shown in Figure 1, these challenges and considerations include:
Emerging trends – the rapid development of AI technologies requires constant adaptation to new innovations in food systems management.
Research gaps – there are still significant gaps in AI research, particularly regarding data availability, model accuracy, and the ability to handle complex food safety challenges.
Ethical considerations – the use of AI in food systems raises concerns about transparency, bias, and data privacy, necessitating responsible and ethical deployment of these technologies.
Economic constraints – implementing AI technologies can be financially challenging, especially for smaller food producers, limiting widespread adoption.
Importance of responsible use – ensuring that AI is used responsibly and with proper oversight is critical to avoid over-reliance on automation and to maintain high food safety standards.
This review examines how AI-based systems are transforming the food supply chain by enabling real-time monitoring, improving traceability, and enhancing overall risk management. Through AI-driven data collection and analysis, food safety measures become proactive, identifying risks before they escalate. Additionally, the integration of AI with advanced food processing methods such as pulsed light treatment and cold plasma creates opportunities for safer and higher-quality food products through precision automation and predictive analytics.
By integrating AI with other emerging technologies such as the IoT and blockchain, food safety management can be further enhanced. These technologies work together to provide continuous monitoring, ensure data transparency, and support quick interventions when food safety risks are detected. For example, IoT-enabled smart sensors can provide real-time updates on temperature and humidity levels, while blockchain can enhance traceability by securely recording every step in the supply chain.
Given the increasing prevalence of foodborne illnesses and the complexity of modern food supply chains, AI offers a path toward a safer, more transparent, and a more resilient food system. By addressing the economic, ethical, and research-related barriers to AI implementation, this review seeks to highlight the transformative potential of AI in food systems, ultimately contributing to improved public health outcomes and greater consumer confidence.
Food safety: protecting public health
Food safety ensures that food products are free from harmful contaminants, pathogens, and toxic substances, safeguarding public health. Despite significant advancements, millions of people are still affected by foodborne diseases annually. Contamination risks occur at multiple stages of the supply chain, necessitating robust detection, monitoring, and intervention mechanisms [2, 3, 4, 5–6]
Governmental agencies like the US Food and Drug Administration (FDA) play a critical role in assessing risks from pathogens, heavy metals, and pesticides. Advanced approaches like mixture toxicology, which evaluates cumulative exposures rather than single contaminants, are increasingly emphasized to enhance consumer protection [4, 5].
AI technologies, including ML and CV, are revolutionizing food safety through:
Contamination detection: AI-powered sensors can detect pathogens like Salmonella and Listeria in real time, enabling swift interventions.
Traceability: Blockchain technologies integrated with IoT devices create an immutable record of the production process, enhancing recall efficiency.
Predictive maintenance: AI analyzes equipment data to predict failures and maintains hygiene, preventing contamination during processing.
The integration of AI-driven systems transforms food safety management from a reactive to a proactive approach, protecting consumers and enhancing trust in global food systems.
Food quality: ensuring consumer satisfaction
Food quality focuses on meeting consumer expectations for taste, texture, appearance, and nutritional value. Maintaining consistency is challenging, especially in globalized supply chains with diverse standards [7, 8,9, 10].
Emerging technologies such as ultraviolet (UV) treatment, high-pressure processing (HPP), and pulsed electric fields (PEF) are redefining quality control by ensuring uniformity and extending shelf life. AI further enhances these techniques by:
Quality inspection: CV systems detect visual defects like discoloration or bruising with unparalleled accuracy, ensuring product consistency.
Shelf-life prediction: ML models analyze storage conditions to predict spoilage, reducing food waste.
Dynamic process control: AI-powered systems monitor real-time data, adjusting processing parameters to maintain quality.
By automating inspections and optimizing production, AI strengthens food quality management, meeting evolving consumer demands while minimizing losses [11, 12, 13, 14, 15, 16, 17, 18, 19–20].
Food security: addressing global challenges
Food security involves ensuring the availability, accessibility, and safety of food for all populations. Challenges such as climate change, urbanization, and resource limitations exacerbate vulnerabilities in the food system [21, 22, 23].
Innovative solutions like food valorization, which converts waste into valuable resources, promote sustainability. However, maintaining the safety and quality of valorized products is critical [25]. AI addresses food security challenges by:
Supply chain optimization: AI models predict disruptions due to extreme weather or logistical issues, enabling proactive adjustments.
Waste reduction: Predictive analytics identify surplus food for redistribution, reducing spoilage.
Precision agriculture: IoT devices and ML optimize crop yields under changing environmental conditions, ensuring resource efficiency.
Through these applications, AI enhances resilience and equity in food systems, supporting global efforts to combat hunger and malnutrition [24, 25].
AI’s role in transforming food systems
AI technologies—ML, NLP, CV, and RL—have emerged as game changers in food systems. By analyzing large datasets in real time, AI provides scalable solutions for detection, monitoring, and risk mitigation [26, 27, 28, 29, 30, 31, 32–33].
AI-driven applications include:
Real-time monitoring: AI systems detect contamination and quality deviations instantly, enabling swift responses.
Proactive risk management: Predictive models identify hazards before they escalate, minimizing risks.
Enhanced traceability: Blockchain-integrated AI ensures transparency throughout the supply chain.
Moreover, combining AI with IoT and blockchain technologies enables seamless data collection and analysis, strengthening food systems’ transparency, efficiency, and security.
Figure 2 presents a comprehensive examination of the many AI-driven methodologies examined in this scholarly review work within the domain of food systems.
Fig. 2 [Images not available. See PDF.]
Comprehensive AI-driven framework for enhancing food systems: from technologies to applications in risk mitigation, monitoring, and traceability
Methodology and literature selection
To provide a thorough and transparent review of the existing literature on AI applications in food systems, particularly in enhancing food safety, quality, and security, we followed a systematic approach based on well-defined inclusion and exclusion criteria. These criteria ensured that only high-quality and relevant studies were considered. The research questions guiding this review helped focus the selection process on key areas where AI plays a role in improving food safety, quality assurance, and food security.
Inclusion and exclusion criteria
Inclusion criteria:
Studies published in peer-reviewed journals or conferences between 1990–2024.
Research aimed at improving food safety, quality, and security, including the detection of foodborne illnesses, compliance monitoring, quality assessment, and enhancing traceability throughout the supply chain.
Studies that applied AI technologies such as ML, NLP, CV and RL to optimize food systems.
Articles written in English, ensuring accessibility and clarity.
Papers providing clear, replicable methodologies and robust results, demonstrating significant impact or potential for real-world application in food safety, quality and security.
Studies covering a range of food types, including agricultural products (e.g., grains, fruits, and vegetables), live animal produce (e.g., chicken, pork, fish, and meat), processed foods, canned goods, and preserved foods. This inclusion ensured a comprehensive overview of the diversity of AI applications across the food industry.
Exclusion criteria:
Studies lacking clear methodologies or sufficient data for replicating results.
Articles published in non-peer-reviewed forums or that exhibited a high degree of bias, thereby compromising the validity of their findings.
Literature search process
The literature search was conducted with the aim of answering the following research questions:
How can AI technologies enhance the detection, monitoring, and prevention of foodborne hazards while improving overall food quality and security?
What are the emerging challenges and ethical considerations when integrating AI into food systems to ensure safety, quality, and security?
What are the most promising AI applications for real-time monitoring and enhancing food safety, quality control, and security standards throughout the food supply chain?
To address these questions, we conducted an extensive search across multiple scientific databases, including PubMed, IEEE Xplore, Web of Science, and Google Scholar. Keywords used for the search included “artificial intelligence,” “machine learning,” “deep learning”, “natural language processing”, “computer vision”, “reinforcement learning”, “food safety,” “compliance monitoring,” and “quality assessment.”
The screening process began with an evaluation of titles and abstracts. Full-text articles meeting the inclusion criteria were reviewed in detail, while duplicates and studies that did not meet the established criteria were excluded. Additionally, studies were assessed for their relevance to the specified food types to ensure the review encompassed a wide spectrum of food categories, including agricultural, animal-derived, processed, and preserved foods. This selection process ensured that the final sources were highly relevant, diverse, and of sufficient quality to provide meaningful contributions to the review's conclusions.
AI technologies in food safety, quality, and security: Key concepts and applications
Figure 3 focuses on the integration of diverse data sources, advanced AI processing techniques, and the generation of outputs like real-time risk alerts, traceability reports, and risk prediction models for effective management of food systems.
Fig. 3 [Images not available. See PDF.]
AI-driven predictive food systems: leveraging data sources, AI processing, and real-time risk management
Definition and key concepts of AI
AI has become indispensable in addressing the complexities of food safety, quality, and security, offering advanced tools for improved detection, prevention, and control across food systems while also revolutionizing the food valorization processes. Beyond detection and monitoring, AI enhances the efficacy of advanced food processing techniques such as high-pressure processing (HPP), ultraviolet (UV) treatment, and pulsed electric fields (PEF), which are pivotal in ensuring food safety and extending shelf life. By harnessing AI technologies, the industry can achieve enhanced monitoring, superior quality control, and proactive interventions, which are critical to maintaining global food standards. Below, we explore the key AI technologies and their collaborative contributions to managing food safety, quality, and security.
ML [34, 35, 36–37] is foundational to the application of AI in food systems and valorization processes, enabling the analysis of vast datasets from environmental sensors, production logs, and historical records to detect patterns and anomalies related to contamination, spoilage, and by-product quality. In advanced food processing, ML algorithms optimize parameters for techniques like high-pressure processing (HPP), predicting ideal conditions to maximize pathogen inactivation while preserving product quality and valorization potential [38, 39]. Similarly, in ultraviolet (UV) treatment, ML dynamically adjusts exposure levels to ensure sterilization efficiency while assessing by-product viability for reuse. Additionally, ML supports innovations in nanotechnology-based packaging [40], which integrates metal-based nanoparticles with packaging materials to enhance physical, chemical, mechanical, and thermal properties, providing superior barriers against gas, moisture, and light while extending shelf life and ensuring safety. By continuously monitoring parameters such as temperature and humidity across the supply chain, ML not only ensures consumer safety but also identifies opportunities to repurpose waste into high-value products, such as dietary fibers or bioactive compounds. Through predictive analytics, ML drives sustainability, enhances packaging durability, mitigates risks, and supports the commercialization of novel technologies like nanoparticle-based active packaging, fostering a circular economy in food systems.
DL [41, 42, 43, 44, 45, 46–47], particularly through Convolutional Neural Networks (CNNs), builds on ML capabilities to analyze visual data, such as detecting physical defects and contaminants in food products. DL models are highly effective in assessing quality by identifying issues like bruising, mold, or irregular textures that may be difficult for the human eye to detect. By ensuring consistency across production batches, DL helps maintain food quality and prevents unsafe products from reaching consumers, thus safeguarding food safety while reinforcing product reliability and consumer confidence.
NLP [48, 49, 50, 51, 52–53] complements ML and DL by enabling the analysis of unstructured text data, such as inspection reports, consumer reviews, and social media posts. NLP extracts insights from these sources to identify emerging risks and quality concerns in real time. For example, NLP can monitor consumer complaints on social media to detect potential foodborne illness outbreaks, allowing authorities to intervene swiftly. This integration enhances decision-making by turning vast amounts of text data into actionable insights, thus improving both food safety and consumer satisfaction by addressing security concerns and quality expectations.
CV [54, 55, 56, 57, 58–59] automates the visual inspection of food products and by-products, leveraging advanced imaging technologies to detect foreign objects, quality defects, or irregularities in shape, texture, and color. In the context of valorization, CV plays a critical role in assessing the quality and suitability of by-products for reuse in creating value-added products, such as dietary fibers or bioactive compounds [60]. By streamlining the inspection process, CV enhances the speed and accuracy of quality assessments, reducing the risk of contaminants or unsuitable materials entering valorization workflows. This not only improves food safety by preventing contamination but also ensures that valorized products meet stringent quality and safety standards, driving sustainability and efficiency in food systems.
Big Data Analytics [61, 62, 63, 64–65] synthesizes information from diverse sources, such as farm practices, environmental conditions, and supply chain operations. This comprehensive analysis enhances food security by identifying contamination risks across the entire supply chain and predicting future safety or quality issues based on historical trends. With this holistic view, stakeholders can implement effective control measures that enhance food safety, improve quality consistency, and secure the food supply from production to consumption.
IoT [66, 67, 68, 69, 70, 71–72] further enhances AI’s role in food systems by facilitating real-time data collection through smart sensors and RFID tags. IoT continuously monitors critical parameters like temperature and humidity throughout the supply chain, ensuring optimal storage conditions that prevent spoilage and contamination. IoT-enabled sensors also monitor real-time conditions during advanced processing techniques like cold plasma and pulsed light treatments, feeding data into AI systems for precise control. For example, IoT sensors detect plasma intensity and adjust parameters to enhance microbial inactivation without compromising product quality. IoT systems also enhance traceability, offering transparency from farm to fork, enabling rapid identification and isolation of contaminated batches, which is vital for both food safety and security.
AI-powered sensors play a key role in real-time contamination detection, including pathogens such as Salmonella or E. coli. These sensors ensure that unsafe products are identified before reaching consumers, reducing the risk of foodborne illness outbreaks. Additionally, sensors detect chemical residues, allergens, and other contaminants, ensuring that food products meet the highest safety and quality standards, thus reinforcing consumer trust and protecting public health [73, 74, 75, 76–77].
Robotics and automation, driven by AI, contribute significantly to food valorization, precision agriculture, and the enhancement of food safety and quality by minimizing human contact during production and processing. AI-powered robots are employed in targeted applications such as pesticide spraying, automated harvesting, and precision processing, which not only reduce contamination risks but also improve the efficiency of valorization processes by optimizing the recovery and transformation of by-products into high-value products. In food valorization, AI-enabled robotics play a critical role in sorting, grading, and processing by-products, ensuring consistent quality and safety [78]. Additionally, AI-driven sanitation systems ensure strict adherence to hygiene protocols in processing environments, further minimizing contamination risks and maintaining the integrity of valorized food products. These technologies collectively enhance operational efficiency, reduce waste, and support the production of sustainable, high-quality outputs [79, 80–81].
Predictive maintenance [82], another key AI application, ensures the reliability of food processing equipment by analyzing sensor data to predict when maintenance is required. This approach prevents equipment breakdowns that could compromise food safety or lead to production downtime, enhancing the efficiency and safety of food production systems.
Finally, blockchain technology adds a layer of transparency and security by providing an immutable, decentralized record of transactions across the supply chain. Blockchain enhances food safety and security by enabling traceability from production to consumption, allowing stakeholders to quickly identify contamination sources. This promotes collaboration across the supply chain and builds consumer trust by ensuring the integrity of food safety protocols [83, 84, 85, 86–87].
In conclusion, the integration of AI into food safety, quality, security, and valorization management provides a transformative approach to addressing modern challenges. By harnessing the power of ML, DL, NLP, CV, big data analytics, IoT, blockchain, and robotics, AI not only enhances the detection, prevention, and response to risks but also optimizes the valorization of by-products into high-value resources. This comprehensive application ensures a safer, more sustainable, and reliable food supply chain, meeting the highest standards of quality while promoting circular economy principles and reducing waste.
Types of ML models relevant to food safety, quality, and security
ML technologies play a crucial role in enhancing food systems by providing advanced tools for monitoring, prediction, and risk mitigation. To fully understand how ML addresses challenges in each domain, it is important to distinguish the distinct roles of food safety, food quality, and food security.
ML applications for food safety and quality control
ML plays a pivotal role in enhancing food safety, quality, and security by providing predictive and automated solutions to traditionally labor-intensive processes. By analyzing vast amounts of data, ML models can identify patterns and anomalies that indicate contamination or spoilage, which is crucial for maintaining food safety standards and reducing the risk of foodborne illnesses. These models enable real-time monitoring and early detection of potential hazards, allowing for swift intervention and prevention. Additionally, ML optimizes supply chain logistics by ensuring that perishable goods are stored and transported under optimal conditions, thereby preserving food quality and preventing spoilage [88]. Beyond these applications, ML also contributes to food security by analyzing historical data and consumer feedback to predict and prevent pathogen outbreaks, ensuring a reliable supply of safe and high-quality food products. Through these comprehensive applications, ML plays a vital role in modernizing and safeguarding global food systems.
Supervised learning: predictive models for food safety and quality
Supervised learning plays a critical role in addressing the three primary concerns of food safety, quality, and security by providing data-driven, predictive solutions across the entire food supply chain. For food safety, models like Support Vector Machines (SVM) and Decision Trees are highly effective at identifying patterns that indicate contamination, such as pesticide residues and microbial growth. This enables early detection of hazards, helping to prevent foodborne illnesses and ensuring a safer food supply [89, 90, 91, 92–93]. When it comes to food quality, supervised learning models like Multiple Linear Regression (MLR) are used to predict shelf life and monitor quality control by analyzing transportation and storage conditions. These predictions allow companies to mitigate spoilage risks and maintain the integrity of perishable items [94, 95]. Supervised models, including Random Forests and SVMs, are also applied to detect defects and anomalies in food products, ensuring that only high-quality items reach consumers [96, 97–98].
In terms of food security, time series analysis through supervised learning frameworks is used to monitor supply chain conditions, ensuring that disruptions are detected and addressed before they impact food availability. This predictive approach guarantees the smooth operation of supply chains, which is essential for maintaining a consistent and secure food supply [99]. Additionally, models like Logistic Regression and SVMs are employed to predict pathogen outbreaks, using historical and environmental data to forecast and prevent the spread of foodborne pathogens, further strengthening food security measures [100, 101, 102–103]. This holistic application of supervised learning enhances food systems by providing effective solutions to prevent contamination, ensure product quality, and secure the food supply chain.
A detailed breakdown of the different uses of supervised ML models in food contamination detection and quality control have been shown in Table 1.
Table 1. Application of supervised ML models for predictive modeling in food contamination detection and quality control
Application | Type of ML model | Benefits | References |
|---|---|---|---|
Contaminant detection | Classification (SVM, decision trees) | Detect pesticide residues and microbial contamination | [84, 85, 86, 87–88] |
Shelf life prediction | Regression (MLR) | Predict the shelf life of perishable items | [89, 90] |
Quality control | Classification (random forests, SVM) | Detect defects and anomalies in food products | [91, 92–93] |
Supply chain monitoring | Time-series analysis | Monitor and predict deviations to optimize conditions | [94] |
Pathogen outbreak prediction | Classification (logistic regression, SVM) | Predict likelihood of outbreaks using historical and environmental data | [95, 96, 97–98] |
Unsupervised learning: detecting anomalies in food safety and quality
Unsupervised learning enhances food safety by uncovering hidden patterns and relationships in large, complex datasets without the need for prior labeling [104]. This approach is critical for anomaly detection, particularly in identifying food contamination and spoilage—key factors in maintaining food safety. Clustering algorithms like K-means [105] can detect unusual data points that signal potential contamination or spoilage. Additionally, unsupervised methods like Principal Component Analysis (PCA) [106] support food quality control by clustering sensor data from food storage environments and identifying deviations from normal conditions. Unsupervised learning also contributes to food security by analyzing vast datasets to predict supply chain risks.
Clustering and anomaly detection in food safety monitoring.
Clustering and anomaly detection are essential techniques in food safety as they help identify patterns, detect anomalies, and manage large datasets critical for ensuring food quality and safety throughout the supply chain. These techniques are used to address various aspects of food safety, from microbial profiling to supply chain monitoring.
Microbial profiling: Clustering algorithms such as K-means and Hierarchical Clustering have been applied to identify typical microbial compositions in food products, helping to detect microbial contamination early. In a recent study [107, 108], K-means was used to cluster microbial profiles in dairy products, successfully identifying bacterial contamination with a 95% accuracy rate, leading to enhanced monitoring protocols in the dairy industry.
Shelf life prediction: K-means and DBSCAN clustering methods are used to group food items with similar storage conditions and spoilage patterns. For instance, a study on perishable foods [107, 108] utilized DBSCAN to predict spoilage risks based on temperature and humidity data. The model successfully optimized storage conditions, reducing spoilage by a significant percentage over six months.
Consumer feedback analysis: By applying K-means and Agglomerative Clustering, companies can cluster consumer feedback to identify common issues related to food safety and quality. A notable example is a project [107, 108] where these techniques were applied to multiple online reviews, leading to the identification of recurring concerns about allergen mislabeling. This resulted in improved labeling protocols and a significant increase in consumer trust.
Contaminant detection: Anomaly detection algorithms such as One-Class SVM and Isolation Forest are used to identify outliers that indicate unusual levels of contaminants. In one study, One-Class SVM was able to detect pesticide residue anomalies in fruit samples with around 90% success rate [109], helping authorities to isolate contaminated batches before distribution.
Supply chain monitoring: Isolation Forest and Autoencoders are applied to monitor transportation conditions in the supply chain. For example, a study monitoring cold-chain logistics for seafood [110] used Isolation Forest to detect deviations in temperature that led to a 30% reduction in spoilage during transportation.
Quality control: Anomaly detection methods like Gaussian Mixture Models are widely used for detecting defects or deviations in production parameters. In a beverage production line [111], these models helped identify equipment malfunctions affecting product consistency, leading to significant reduction in defective products.
Predictive models for microbial behavior: Clustering models such as K-means are used to understand microbial behavior, which aids in designing safety interventions. In one case [112], K-means clustering of microbial growth patterns in ready-to-eat foods helped create more effective storage and packaging interventions, reducing contamination by 25%.
Temperature-sensitive goods monitoring: Isolation Forest has been used to monitor temperature-sensitive goods during storage and transportation. A case study [110] demonstrated that the algorithm effectively detected anomalies in storage conditions, reducing spoilage by a significant amount across a short period.
Table 2. Applications of clustering models for anomaly detection in food safety
Application | Type of ML model | Benefits | References |
|---|---|---|---|
Microbial profiling | Clustering (K-means, Hierarchical Clustering) | Identify typical microbial compositions and detect contamination | [107, 108] |
Shelf life prediction | Clustering (K-means, DBSCAN) | Group food items with similar storage conditions and spoilage patterns | [107, 108] |
Consumer feedback analysis | Clustering (K-means, Agglomerative Clustering) | Identify common issues related to food safety and quality | [107, 108] |
Contaminant detection | Anomaly Detection (One-Class SVM, Isolation Forest) | Identify outliers indicating unusual levels of contaminants | [109] |
Supply chain monitoring | Anomaly detection (isolation forest, autoencoders) | Detect irregularities in transportation conditions | [110] |
Quality control | Anomaly detection (isolation forest, Gaussian mixture models) | Identify defects or deviations in production parameters | [111] |
Predictive models for microbial behavior | Clustering (K-means) | Understand microbial behavior to design safety interventions | [112] |
Temperature-sensitive goods monitoring | Anomaly detection (isolation forest) | Optimize storage and transportation to reduce spoilage risks | [110] |
Pattern recognition in large datasets for risk assessment in food systems.
Effective risk assessment in food safety relies on identifying patterns in large datasets. Advanced unsupervised learning techniques such as clustering and anomaly detection provide insights that are crucial for preventing foodborne illness outbreaks and ensuring compliance with safety standards.
Foodborne illness outbreak detection: DBSCAN is a clustering method commonly used to detect clusters of foodborne illness cases. A significant study [113] applied DBSCAN to health data from several regions, leading to early identification of a foodborne illness outbreak, which reduced the impact of the outbreak by enabling faster public health responses.
Consumer complaint analysis: K-means and Agglomerative Clustering are used to cluster consumer complaints and identify recurring safety issues. In a study of 100,000 complaints [114, 115], these models helped isolate major issues related to food contamination, enabling food producers to take corrective actions, reducing recalls by over 10%.
The applications and benefits of clustering models for risk assessment in food systems are stated in Table 3.
Table 3. Applications of clustering models for risk assessment in food systems
Application
Type of ML model
Benefits
References
Foodborne illness outbreak detection
Clustering (DBSCAN)
Identify clusters of foodborne illness cases
[113]
Consumer complaint analysis
Clustering (K-means, agglomerative clustering)
Identify and address common safety and quality issues
[114, 115]
RL for food safety, quality and security
RL [116, 117] plays a crucial role in ensuring food safety, quality, and security by optimizing decision-making in dynamic, complex environments such as food processing and supply chains. RL involves agents learning from trial and error to achieve optimal outcomes, which is highly beneficial for automating processes and improving efficiency in the food industry. This allows systems to continually adapt and improve their strategies for maintaining food safety standards, enhancing quality control, and optimizing food security [117].
Enhancing food safety with RL
In terms of food safety, RL models like Q-Learning and Deep Q-Networks (DQN) have proven instrumental in improving compliance monitoring and contamination detection. For example, RL algorithms monitor processing machinery in real-time, detecting potential failures or contamination risks before they compromise food safety. Predictive maintenance, enabled by RL, reduces the risk of foodborne illnesses by ensuring equipment remains functional and clean, avoiding breakdowns that could lead to contamination [118]. Additionally, RL-powered training platforms help food safety staff practice handling contamination scenarios, improving their ability to respond effectively and adhere to safety protocols. These systems have demonstrated a 30% improvement in response times and a 15% increase in compliance with food safety regulations [118, 123]. Moreover, RL models assist in automating food recall processes, identifying contaminated products in real-time, minimizing public health risks, and reducing the overall impact of food recalls by 25% [118].
Improving food quality with RL
RL also plays a critical role in maintaining food quality by automating inspection processes and optimizing equipment functionality over time. By employing Deep Q-Networks (DQN) and Policy Gradient Methods, RL systems can monitor and detect physical defects in food products, such as bruising or contamination, ensuring consistent quality across batches [118]. RL models further optimize supply chain processes, improving storage conditions and reducing spoilage risks, particularly in temperature-sensitive environments like cold storage. In a cold storage study, RL models reduced energy consumption by 15% and increased product shelf life by 20%, ensuring that high-quality products are delivered to consumers [118, 119, 120, 121–122]. By improving predictive maintenance systems and reducing unplanned downtime, RL systems also ensure that food production remains efficient and contamination-free, resulting in better product quality and reliability [118].
Securing food supply chains with RL
For food security, RL models excel at optimizing supply chain logistics and ensuring that perishable items are transported and stored under ideal conditions. By applying Q-Learning and Deep Q-Networks (DQN) to cold chain management, RL algorithms dynamically adjust transportation routes and storage conditions, maintaining safe temperatures throughout transit. This approach has been shown to reduce spoilage by 20% and enhance overall supply chain efficiency, ensuring a more secure and reliable flow of goods [124]. Furthermore, RL’s predictive capabilities extend to consumer safety and preferences, where models can analyze dietary restrictions and personalize recommendations to ensure consumers select safe and suitable products. This not only increases consumer satisfaction but also reduces the risks associated with allergens or other unsafe food choices [118]. Additionally, smart kitchen appliances powered by RL adjust cooking parameters in real-time, ensuring consistent cooking results and reducing the risk of undercooked or unsafe food items [125].
A comprehensive overview of the use cases, descriptions, types of RL models used, and references in the context of RL applications in food safety, quality and security have been shown in Table 4.
Table 4. Application of RL algorithms in food safety, quality and security
Application | Type of RL model | Benefits | References |
|---|---|---|---|
Process optimization | Q-Learning, DQN | Optimization of supply chain routes and storage conditions, and critical manufacturing processes | [121] |
Quality control | DQN, Policy Gradient methods | Enhancement of automated inspection systems and predictive maintenance for machinery | [118] |
Food storage and preservation | Q-learning, DQN | Optimization of refrigeration systems and shelf-life prediction | [118] |
Safety protocols and compliance | PPO, DQN | Training staff through simulation environments and ensuring regulatory compliance | [118, 123] |
Consumer safety and preferences | Q-learning, DQN | Personalized recommendations and efficient food recall processes | [118] |
Cold chain management | Q-learning, DQN | Optimization of cold chain logistics for perishable goods | [124] |
Smart kitchens | DQN, Policy Gradient methods | Development of smart kitchen appliances that adjust cooking parameters in real-time | [125] |
DL for advancing food safety, quality, and security
DL, a subset of AI, utilizes neural networks with multiple layers to model and understand complex patterns in data. DL, particularly through neural networks, has made significant strides in food safety, quality, and security. These models, which include CNNs and Recurrent Neural Networks (RNNs), are highly effective at processing large and complex datasets, making them invaluable tools for monitoring, detecting, and predicting food-related risks across the supply chain.
One significant application of DL in food safety is in the detection and prediction of foodborne illness outbreaks. CNNs [126] and RNNs [127] are employed to analyze large datasets from sources such as social media, healthcare records, and sensor data. By identifying patterns and anomalies, these models can detect and predict outbreaks early, allowing for prompt intervention and mitigation efforts.
Ingredient similarity analysis is another crucial area where DL makes a substantial impact. Autoencoders and CNNs analyze the chemical compositions and properties of ingredients, categorizing and identifying similarities. This ensures the use of safe and compatible ingredients in food production, reducing the risk of adverse reactions and contamination.
In supply chain segmentation, Long Short-Term Memory (LSTM) networks and CNNs play a pivotal role. These models segment various stages of the supply chain to identify and mitigate contamination risks. By analyzing data from different points in the supply chain, they help ensure that each stage adheres to safety standards, maintaining the overall integrity of the food supply [128].
Real-time quality control is significantly enhanced by DL models like CNNs and Generative Adversarial Networks (GANs) [129]. These models monitor production data in real-time to detect defects and contamination in food products. Immediate corrective actions can be taken based on these detections, ensuring consistent quality and safety of the food products reaching consumers.
Environmental monitoring is another critical application of DL in food safety. CNNs and RNNs analyze environmental data, such as temperature, humidity, and contamination levels, to prevent spoilage and ensure optimal storage conditions. These models help maintain the safety and quality of food during storage and transportation, reducing waste and ensuring consumer safety.
Predictive maintenance of equipment is facilitated by DL models like CNNs and LSTM networks. By analyzing operational data, these models predict equipment failures, enabling timely maintenance and preventing breakdowns. This proactive approach maintains hygiene standards and reduces the risk of contamination due to equipment malfunctions.
DL models also contribute to understanding microbial behavior in food processing environments. Models like CNNs and RNNs predict microbial behavior under various conditions, assisting in designing effective safety interventions. This helps ensure that food processing environments remain free from harmful microbial activity, enhancing overall food safety.
Lastly, DL enhances consumer complaint analysis. LSTM networks and CNNs analyze consumer complaints to identify common safety and quality issues. Understanding these patterns allows food companies to address recurring problems, improving overall food safety and consumer satisfaction.
DL is revolutionizing food safety by providing advanced tools for detecting, predicting, and mitigating risks throughout the food supply chain. These technologies enhance the ability to monitor and control food quality, ensuring that consumers receive safe and high-quality products.
CNNs for visual data analysis in food safety
CNNs are a specialized class of DL models particularly well-suited for tasks involving spatial data and pattern recognition. In the context of food safety, CNNs play a critical role due to their ability to automatically and efficiently analyze and interpret complex visual and sensor data. Their importance and applications in food safety are vast and varied, encompassing several key areas.
CNNs are essential for ensuring food safety through visual data analysis. They are particularly useful for identifying contaminants, defects, or irregularities that might not be easily detectable by the human eye. By automating visual inspection processes, CNNs improve food quality control and reduce the risk of human error during the production phase. These systems can detect subtle changes in product appearance, such as color or texture deviations, ensuring that only high-quality products reach the consumer [130]. In terms of food security, CNNs also aid in the classification of food products and packaging, ensuring accurate inventory management and preventing fraudulent practices, such as food mislabeling [131, 132].
A detailed breakdown of the different uses of CNNs in food safety have been shown in Table 5. It provides a broader overview of the diverse applications of CNNs in food safety, from quality control and fraud detection to environmental monitoring and microbial contamination detection.
Table 5. Application of CNN algorithms in food systems
Application | Benefits | References |
|---|---|---|
Quality control and defect detection | CNNs analyze images of food products to identify defects such as blemishes, discoloration, or foreign objects | [130] |
Food fraud detection | CNNs verify the authenticity of food products by analyzing packaging and labels to detect counterfeiting and mislabeling | [131, 132] |
Contamination detection | CNNs monitor food processing environments by analyzing images or sensor data to identify signs of contamination such as spillage or foreign objects | [133] |
Environmental monitoring | CNNs analyze images and sensor data related to environmental conditions to ensure optimal storage and prevent spoilage | [135] |
Food classification and identification | CNNs classify and identify different types of food products based on visual characteristics, improving sorting and inventory management | [134] |
Texture analysis | CNNs analyze food texture from images to ensure consistency and quality, identifying issues like irregularities or undesired textures | [133] |
Ingredient verification | CNNs verify the presence and quality of ingredients by analyzing images, ensuring that only the correct and high-quality ingredients are used | [135] |
Label and packaging inspection | CNNs inspect food labels and packaging for accuracy and compliance with regulations, detecting errors or fraudulent practices | [135] |
Microbial contamination detection | CNNs analyze images of microbial growth or contamination to detect and quantify microbial presence in food products | [136] |
RNNs for time-series analysis in food systems
RNNs are powerful tools in food safety, quality, and security due to their ability to process sequential data and capture temporal dependencies. They are particularly effective in predicting contamination risks, monitoring storage conditions, and tracking products throughout the supply chain. RNNs analyze historical data on environmental factors such as temperature, humidity, and contamination incidents, providing accurate predictions about food spoilage and quality degradation. This predictive capability enables better decision-making, improving both food safety and product quality by addressing potential issues early, whether through real-time monitoring or predictive maintenance of equipments [137, 138–139].
RNNs are also used in supply chain monitoring to ensure product integrity from farm to fork. By processing continuous data from sensors and tracking devices, they can detect anomalies such as equipment malfunctions or temperature deviations that could compromise food safety. Their ability to predict supply chain disruptions improves inventory management, reduces waste, and helps maintain consistent food security [139].
Furthermore, RNNs can analyze consumer feedback and public health data to identify trends related to foodborne illnesses, providing early warnings to mitigate potential outbreaks. This real-time analysis contributes to more effective monitoring of the entire food supply chain, ensuring compliance with safety standards and enhancing overall traceability and visibility [141, 142].
Time-series analysis with RNNs for shelf-life prediction and supply chain monitoring
RNNs excel in shelf-life prediction by analyzing time-series data on temperature, humidity, and other environmental conditions that affect the longevity of perishable food items. This helps optimize storage conditions, ensuring products remain fresh for consumers [137]. In supply chain monitoring, RNNs detect real-time anomalies and deviations, maintaining product quality and compliance with safety standards while reducing spoilage and enhancing traceability. By improving logistics and inventory management, RNNs contribute to a more resilient and efficient food supply chain [138, 139, 140, 141–142].
The applications of RNNs in time-series analysis for shelf-life prediction and supply chain monitoring are detailed in Table 6.
Table 6. Application of RNN algorithms for shelf life prediction and supply chain monitoring
Application | Reason for use | Benefits | Challenges | References |
|---|---|---|---|---|
Shelf-life prediction | RNNs analyze time-series data (e.g., temperature, humidity) to predict when perishable food items are likely to spoil or degrade in quality | Optimizes storage conditions. Ensures freshness of products for consumers | Requires high-quality, clean, and relevant data. Understanding complex model decisions | [137] |
Supply chain monitoring | RNNs process continuous data from sensors and tracking devices to monitor the movement and condition of food products throughout the supply chain | Detects anomalies or deviations in real-time. Maintains product integrity. Ensures compliance with safety standards. Enhances traceability | Significant effort needed for seamless integration into existing systems. Complex nature of models may pose interpretability challenges | [138, 139] |
Inventory management | RNNs predict demand and supply trends by analyzing historical sales and demand data | Optimizes logistics. Reduces waste. Minimizes disruptions | Requires continuous monitoring and updating of models. Depends on the accuracy of historical data | [140] |
Real-time insights | RNNs provide real-time insights by processing continuous data from various sources | Timely decision-making. Mitigates risks. Ensures efficient delivery | High computational resources are needed for real-time processing. Potential issues with data synchronization from different sources | [141, 142] |
Enhanced visibility | RNNs enhance visibility and traceability in the supply chain by monitoring and analyzing time-series data from various points in the supply chain | Better inventory management. Improved supply chain resilience | Integration challenges with existing systems. Maintaining data consistency and accuracy throughout the supply chain | [142] |
RNNs for tracking and predicting foodborne illness outbreaks
RNNs are instrumental in tracking foodborne illness outbreaks by analyzing time-series data from various sources such as health records, social media, and public health reports. These networks can identify patterns and trends in the data, enabling the early detection of potential outbreaks. By continuously monitoring and processing this information, RNNs can predict the spread of illness and identify potential sources of contamination. This proactive approach allows public health officials to respond more swiftly and effectively, mitigating the impact of outbreaks. Furthermore, RNNs can help in pinpointing geographic hotspots and vulnerable populations, thereby optimizing resource allocation and intervention strategies to prevent further spread and ensure food safety [143].
NLP for enhancing food safety, quality, and security
NLP significantly enhances food safety, quality, and security by automating the analysis of large volumes of textual data. By processing unstructured information from sources such as health reports, social media posts, consumer reviews, and regulatory documents, NLP enables authorities and food producers to monitor and address safety concerns in real time. This early detection allows for rapid responses to emerging foodborne risks and helps ensure compliance with safety standards throughout the supply chain [144, 145].
NLP contributes to food safety by identifying potential risks from consumer complaints, such as symptoms of foodborne illnesses, and allows businesses to act quickly to mitigate these issues. For food quality, NLP helps businesses monitor product characteristics like taste, texture, and packaging through the analysis of feedback, ensuring that products meet consumer expectations and regulatory standards. Additionally, NLP plays a vital role in food security by assessing and monitoring data from across the supply chain, identifying possible disruptions or contamination risks that could compromise the availability and safety of food products [144, 145–146].
Text analysis for monitoring compliance and risk management
NLP automates the examination of regulatory documents, inspection reports, and compliance records, allowing for efficient identification of violations and non-compliance. It helps companies adhere to food safety regulations by extracting and interpreting critical information from complex texts, ensuring that food processors and suppliers meet the necessary safety standards [143]. NLP further supports quality control by analyzing communication between entities in the supply chain, verifying that food safety and quality standards are maintained.
Additionally, NLP helps keep organizations updated with the latest regulations by summarizing key points from newly published guidelines, scientific research, and regulatory changes. This ensures that businesses stay compliant with evolving safety protocols and are better equipped to maintain high food quality standards [146]. Furthermore, NLP’s ability to process consumer feedback across multiple platforms helps detect safety issues early, leading to quicker resolutions and enhanced consumer trust [147, 148].
The applications of NLP algorithms analysis for textual analysis in monitoring compliance and enhancing food safety are detailed in Table 7.
Table 7. Application of NLP algorithms for textual analysis in monitoring compliance with food safety regulations
Application | Reason for use | Benefits | Challenges | References |
|---|---|---|---|---|
Automated compliance monitoring | NLP algorithms analyze regulatory documents, inspection reports, and compliance records to detect non-compliance | Effectively identifies violations, reduces manual labor, and offers comprehensive compliance monitoring | Requires accurate and up-to-date data for training. Complex legal language may reduce accuracy | [144] |
Document analysis | Extracts and interprets critical information from regulatory texts and stakeholder communications to ensure standards compliance | Automates the extraction of key information, ensuring adherence to safety standards and improving compliance tracking | Extracting relevant data from complex texts can be difficult; models require frequent updates | [145, 146] |
Consumer feedback monitoring | Analyzes consumer feedback from social media, online reviews, and customer service reports to identify potential food safety concerns | Enables early detection of issues, helps in resolving concerns before escalation, and enhances consumer safety | Managing large data volumes and variability in language and sentiment expression poses challenges | [147] |
Regulatory updates | Summarizes and highlights key points from new regulatory documents and research papers to keep industry professionals informed | Keeps organizations current with changing regulations, reduces manual review time, and ensures compliance | Rapidly evolving regulations make it difficult to maintain accuracy and timeliness | [148] |
Supply chain communication | Monitors communication between supply chain entities to ensure adherence to safety protocols and agreements | Verifies compliance with safety protocols and detects non-compliance in supply chain communications | Variations in communication styles and integrating NLP into existing systems can be challenging | [149] |
Sentiment analysis of consumer feedback for early detection and quality control
NLP for sentiment analysis of consumer reviews and feedback is a powerful tool for monitoring and improving food safety. By analyzing the sentiment expressed in customer reviews and feedback, NLP can identify patterns related to food quality, safety concerns, and customer satisfaction. This involves processing and interpreting large volumes of textual data from sources like online reviews, social media, and customer service interactions [150, 151].
For food quality, sentiment analysis identifies consumer perceptions related to taste, texture, and packaging, enabling businesses to improve product quality based on customer feedback. Positive sentiment reinforces good practices, while negative sentiment highlights areas for improvement in quality control.
Regarding food security, sentiment analysis provides insights into the availability and affordability of food products. By understanding consumer feedback on pricing and access to food, companies and regulatory bodies can address potential barriers to ensuring that food is both safe and accessible to all consumers.
Sentiment analysis uses NLP algorithms to categorize text into positive, negative, or neutral sentiments [152]. This helps in detecting emerging issues such as complaints about foodborne illnesses, dissatisfaction with product quality, or concerns about safety practices. By aggregating and analyzing this data, companies can pinpoint common problems, assess the effectiveness of their food safety measures, and make informed decisions to enhance product quality and consumer trust.
By using sentiment analysis, companies can identify recurring problems, evaluate the effectiveness of their safety measures, and make timely decisions to address food safety concerns. For example, a study analyzing social media posts during a foodborne illness outbreak [153, 154] used NLP algorithms to detect increasing complaints about stomach discomfort associated with a particular product. This early warning led to a rapid recall, which minimized public health risks and reduced the spread of the illness.
Moreover, sentiment analysis not only identifies issues but also highlights successful practices. Positive reviews about the consistency of safety standards or packaging integrity can reinforce the company’s best practices, boosting brand reputation and consumer trust.
The applications of NLP algorithms analysis for sentiment analysis of consumer reviews and feedback are detailed in Table 8.
Table 8. Application of NLP algorithms for sentiment analysis of consumer reviews and feedback
Application | Reason for use | Benefits | Challenges | References |
|---|---|---|---|---|
Sentiment analysis | NLP algorithms categorize consumer reviews and feedback into positive, negative, or neutral sentiments to gauge perceptions | Detects patterns in food quality and safety concerns, offers insights into customer satisfaction, and identifies emerging issues early | Requires large volumes of text data for accuracy. Variability in language and sentiment expression may affect results | [153, 154] |
Issue detection | Analyzes consumer sentiment to identify and address reported food safety issues | Facilitates early detection of safety problems, allowing timely interventions and resolutions | Difficult to pinpoint specific issues from generalized feedback. Ensuring sentiment interpretation accuracy is essential | [155] |
Quality improvement | Aggregates sentiment data to assess the effectiveness of food safety measures and product quality | Highlights successful practices and provides actionable insights for improving quality and safety | Continuous monitoring and analysis are required. Processing large datasets can be resource-intensive | [156, 157] |
Consumer satisfaction | Measures overall consumer satisfaction and identifies areas for improvement | Enhances customer trust and loyalty, improving brand reputation by addressing key concerns | Managing diverse feedback from various customer segments. Balancing negative feedback with positive practices | [158] |
Early warning system | Uses sentiment analysis to provide proactive warnings about potential contamination or quality issues | Helps prevent major outbreaks or recalls by addressing potential food safety issues early | Requires accurate and timely interpretation of sentiment trends. Integrating findings into operational processes can be challenging | [140] |
CV for enhancing food safety, quality and security
CV, a field of AI, plays a crucial role in improving food safety, quality, and security by automating visual inspections and monitoring processes across the food industry. By leveraging advanced imaging techniques and ML algorithms, CV systems can detect contaminants, defects, and quality issues in food products, significantly enhancing the accuracy and speed of inspections [159, 160]. For example, CV is employed on production lines to inspect food items like fruits and vegetables for spoilage, foreign objects, or irregularities in shape and color, ensuring that only safe, high-quality products reach consumers [162, 163, 164–165]. This automated approach minimizes human error, reduces labor costs, and improves overall efficiency.
In addition to inspection tasks, CV is integral to maintaining hygiene and safety compliance in food processing environments. Cameras equipped with CV capabilities can continuously scan facilities to ensure cleanliness and adherence to employee safety protocols, such as wearing protective gear, thereby reducing the risk of contamination [167]. CV systems also play a critical role in documenting and tracking every step of the production process, providing the traceability and accountability necessary for identifying potential safety issues and responding promptly when problems arise [169].
Automated inspection and quality assessment using CV
CV significantly enhances automated inspection and quality control by employing advanced imaging techniques and ML algorithms to analyze visual data in real-time. On production lines, CV systems detect defects, contaminants, and anomalies in food items with high precision and speed. For example, these systems inspect produce for signs of spoilage, foreign objects, or imperfections in shape and color, ensuring that only products meeting the highest standards are distributed. This automation improves consistency, reduces waste, and increases productivity by minimizing the chances of human error and lowering labor costs [162, 163].
CV also plays a crucial role in ensuring the quality and security of food products throughout the supply chain. By monitoring each stage of production, packaging, and distribution, CV systems help maintain food safety standards, ensuring compliance with regulatory requirements and enhancing traceability. The ability of CV to provide real-time feedback allows operators to address issues immediately, ensuring that food products meet both safety and quality standards at every point [168].
In addition, CV technology is employed to monitor packaging and labeling accuracy, identifying errors such as incorrect labels, damaged packaging, or missing information. This helps ensure compliance with food safety regulations and protects consumer trust by minimizing the likelihood of mislabeled or unsafe products entering the market [166]. Integrating CV into the food safety management system not only enhances operational efficiency but also builds consumer trust by consistently delivering safe, high-quality food products [161, 169].
The applications of CV algorithms for automated inspection and quality assessment of food products are detailed in Table 9.
Table 9. Application of CV algorithms for automated inspection and quality assessment of food products
Application | Reason for use | Benefits | References |
|---|---|---|---|
Defect detection | Identifies spoilage, contaminants, and defects in food items based on shape, color, and texture | Ensures product quality, reduces waste, and minimizes human error | [162, 163] |
Foreign object detection | Detects foreign objects in food products during production | Enhances food safety, prevents contamination, and protects consumer health | [164, 165] |
Packaging and labeling accuracy | Inspects packaging and labeling for correctness, including label content, placement, and packaging integrity | Ensures regulatory compliance, reduces packaging errors, and maintains brand reputation | [166] |
Hygiene and safety compliance monitoring | Monitors processing facilities to ensure cleanliness and adherence to safety protocols by employees | Maintains high hygiene standards, reduces contamination risk, and enhances worker safety | [167] |
Real-time feedback and adjustment | Provides instant feedback to operators for correcting issues during the production process | Increases operational efficiency, allows immediate adjustments, and minimizes production downtime | [168] |
Traceability and documentation | Tracks and documents each step of the food production process for accountability and traceability | Enhances traceability, aids in recall management, and improves transparency | [169] |
Real-time monitoring and compliance in food production lines
CV plays a vital role in real-time monitoring of food production by utilizing advanced imaging techniques and ML algorithms to detect contaminants, defects, and deviations in food products, enabling immediate corrective actions. This ensures that only high-quality products proceed to packaging and distribution, improving consumer safety and reducing waste [170, 171]. Additionally, CV monitors the overall production environment for hygiene and safety compliance, analyzing real-time images to ensure cleanliness, proper equipment operation, and adherence to safety protocols. This real-time data enhances traceability, making it easier to address any issues in the production process.
Data sources and data management
Figure 4 highlights critical challenges in AI implementation across food safety, quality, and security, emphasizing the need to address data privacy, bias in AI models, and transparency issues. These challenges are heightened by the diverse data sources involved in ensuring food safety, maintaining quality standards, and securing the food supply chain. Data sources in this field range from regulatory documents and inspection reports to environmental sensor data, production line images, and consumer feedback, all of which provide vital insights into the state of food systems.
Fig. 4 [Images not available. See PDF.]
Challenges in AI implementation for food safety: addressing data privacy, bias, and transparency issues
Regulatory documents and inspection reports deliver structured data on compliance with safety standards, while environmental sensors, such as temperature and humidity monitors, provide real-time data crucial for maintaining optimal storage conditions and preventing spoilage [172, 173]. Images and videos captured from production lines, processed through CV systems, improve quality control by providing detailed visual data on potential contaminants, defects, and other safety issues. Furthermore, unstructured data from consumer feedback, sourced from platforms such as social media, online reviews, and customer service interactions, is invaluable in identifying emerging food safety issues and trends.
Effectively managing this wide array of data requires robust data management systems and practices. Data from these multiple sources must be integrated to provide a holistic view of food safety throughout the supply chain. This integration is essential for enabling accurate and timely decision-making. Advanced analytics techniques, such as ML and NLP, play a key role in extracting actionable insights from both structured and unstructured data. In addition, data governance policies must ensure data privacy and security, maintain regulatory compliance, and protect consumer trust. Continuous monitoring and regular updates to data management systems are also vital for keeping pace with evolving food safety standards and new technological advancements, thereby enhancing traceability, accountability, and decision-making across the food industry.
Types of data
In food safety management, several types of data are collected and analyzed to ensure food quality, safety, and security. These include sensory data, images, and textual data, each contributing to different aspects of food safety management.
Sensory Data: Sensory data, such as temperature, humidity, and gas levels, are collected through sensors placed in production and storage environments. This data is crucial for maintaining optimal conditions that prevent spoilage and contamination. For instance, temperature sensors in refrigerated storage units ensure that perishable goods remain within safe temperature ranges, while humidity sensors prevent mold growth. Gas sensors can detect harmful levels of gases like carbon dioxide or ammonia, which could indicate spoilage or contamination. Continuous monitoring of these sensory parameters allows for real-time adjustments and interventions to maintain food safety standards [174, 175].
Images: Visual data, captured through cameras and imaging systems, is extensively used in food safety inspections and quality control. CV technologies analyze images and videos from production lines to detect defects, contaminants, and deviations in food products. This can include identifying foreign objects, ensuring proper labeling and packaging, and verifying the visual quality of products such as detecting bruises on fruits or irregular shapes in baked goods. High-resolution images enable detailed inspection, reducing human error and increasing the efficiency and accuracy of the quality control processes [176, 177].
Textual Data: Textual data encompasses a wide range of information, including regulatory documents, inspection reports, scientific research papers, and consumer feedback. NLP algorithms are employed to analyze this data for compliance monitoring, trend analysis, and early detection of potential food safety issues. Regulatory documents and inspection reports provide structured data about safety standards and compliance, while consumer feedback from social media and online reviews offers unstructured data that can reveal emerging food safety concerns. Analyzing textual data helps in staying updated with regulatory changes, understanding consumer perceptions, and making informed decisions to improve food safety practices [178, 179].
Each of these data types—sensory data, images, and textual data—plays a vital role in creating a comprehensive food safety management system. By integrating and analyzing these diverse data sources, food safety professionals can ensure the highest standards of quality and safety throughout the food supply chain.
Methods of data collection and integration
Efficient methods of data collection and integration are crucial for establishing a comprehensive and effective food safety management system. These methods ensure that data from various sources—such as sensors, imaging devices, and textual reports—are accurately collected, stored, and integrated for real-time decision-making.
One common approach is the use of Centralized Data Repositories, which aggregate data from various sources into a structured, and accessible repository. This method allows for easy querying and analysis, enabling food safety professionals to quickly access critical data. However, challenges arise in managing the quality and consistency of data from different sources, as discrepancies in format and accuracy must be standardized to ensure effective integration [180, 181].
Data warehousing is another critical method that integrates data from different operational systems—such as sensory data, images, and textual reports—into a unified repository. This consolidated storage facilitates a holistic view of food safety issues by bringing together diverse data types for comprehensive analysis. However, data warehousing requires careful planning and implementation to ensure compatibility between systems and the seamless flow of data [182, 183].
The ETL (Extract, Transform, Load) process is a widely used method to manage data from disparate sources. ETL involves extracting data, transforming it into a standardized format suitable for analysis, and then loading it into a storage system. This method harmonizes data, making it easier to analyze, but it is often complex to set up and maintain, particularly when dealing with diverse data formats from various sources [184, 185].
Another method, data fusion, combines data from multiple sources to create a comprehensive dataset. This method enhances data quality and completeness, offering a more detailed picture by merging different data types. However, integrating heterogeneous data from various sources can be challenging, especially when the data is inconsistent in format or timing [186, 187].
Cloud-based integration uses cloud platforms to integrate data from multiple sources, offering scalability and flexibility. This approach facilitates real-time analytics and collaboration across different locations, as the data can be accessed from anywhere with an internet connection. Despite the advantages, managing data privacy and security in the cloud remains a significant challenge, particularly when dealing with sensitive food safety information [188, 189].
The different methods of data collection and integration are summarized in Table 10.
Table 10. Methods of data collection and integration in the realm of food systems
Method | Description | Benefits | Challenges | References |
|---|---|---|---|---|
Centralized data repositories | Aggregates data from various sources into a structured repository | Easy access, querying, and analysis | Managing data quality and consistency | [180, 181] |
Data warehousing | Integrates data from different operational systems into a unified repository | Consolidates sensory data, images, and textual data | Requires careful planning and implementation | [182, 183] |
ETL (extract, transform, load) processes | Extracts data from sources, transforms it into a suitable format, and loads it into storage systems | Harmonizes data from disparate sources | Complexity in setting up and maintaining ETL processes | [184, 185] |
Data fusion | Combines data from multiple sources to create a comprehensive dataset | Enhances data quality and completeness | Integrating data from heterogeneous sources | [186, 187] |
Cloud-based integration | Uses cloud platforms for scalable and flexible data integration | Facilitates real-time analytics and collaboration | Managing data privacy and security in the cloud | [188, 189] |
Importance of data preprocessing and cleaning
Data preprocessing and cleaning are essential steps for ensuring the reliability and accuracy of food systems management. These processes transform raw data into formats that can be effectively analyzed, which is crucial for making informed decisions that impact public health, food quality, and supply chain security. Below are the key reasons why data preprocessing and cleaning are vital in managing food systems:
Accuracy and reliability: Raw data often contains errors, duplicates, or inconsistencies that can lead to flawed analysis and poor decision-making. Cleaning the data involves identifying and correcting these issues to ensure the dataset is accurate and reliable. This is particularly important in food systems, where errors in data can result in missed safety violations, quality defects, or contamination, ultimately posing risks to consumers [190].
Improved data quality: Preprocessing techniques such as normalization, transformation, and standardization help improve the quality of the data. High-quality data is critical for robust analysis in food systems, providing a clear and consistent basis for monitoring food safety, quality control, and security across the supply chain. For example, standardized formats and units allow for easier comparisons and trend analysis, ensuring consistency in the data [191, 192].
Enhanced decision-making: Clean and preprocessed data supports better decision-making by providing accurate and actionable insights. Food systems professionals rely on these insights to make important decisions regarding product recalls, regulatory compliance, and improvements in food production and storage processes. Without proper data cleaning and preprocessing, decisions could be based on faulty or incomplete data, leading to inefficient outcomes [193, 194].
Efficient data integration: Preprocessed and cleaned data can be easily integrated from multiple sources, such as sensors, imaging systems, and textual reports. This integration is essential for a comprehensive food systems management approach, allowing seamless cross-referencing and holistic analysis of the supply chain, from production to consumption [195].
Detection of anomalies and trends: Effective data preprocessing enables early detection of anomalies and trends that could signal potential food safety, quality, or security issues. For instance, outliers in sensor data could indicate equipment malfunctions or deviations in storage conditions. Cleaning and preprocessing data enhance the ability to promptly and accurately detect such anomalies [196, 197].
Compliance and reporting: Regulatory compliance requires accurate reporting of food safety and quality data. Preprocessing ensures that data is formatted correctly for regulatory reporting, meeting required standards. Clean data also simplifies the preparation of reports for audits, internal reviews, and regulatory bodies, ensuring transparency and accountability across food systems [198, 199].
Resource optimization: Data cleaning and preprocessing help optimize resources by reducing the time and effort needed for manual data correction and analysis. Automated processes handle large data volumes efficiently, allowing food systems professionals to focus on higher-level strategic initiatives and decision-making [200, 201].
Enhanced ML and AI: Many modern food safety management systems leverage ML and AI for predictive analytics and automated decision-making. These technologies require high-quality, preprocessed data to function effectively. Clean data improves the performance and accuracy of ML models, leading to better predictions and automated insights.
In summary, robust data preprocessing and cleaning practices are essential for maintaining high standards of food safety, quality, and security. These processes ensure that the data used for monitoring and analysis is accurate, reliable, and actionable, enabling better decision-making and improving overall outcomes in food safety management.
AI model development and training
Developing AI models for food safety involves a series of critical steps to ensure the models are accurate, reliable, and effective. The process begins with data collection, which is foundational as the quality and variety of data directly influence model performance [202]. Data is gathered from various sources, including sensors that measure environmental factors, imaging systems that capture product quality, and textual data from regulatory documents and consumer feedback. Ensuring this data is comprehensive and representative of real-world conditions is crucial.
Next is data preprocessing, which involves cleaning and transforming the collected data into a suitable format for analysis. This step includes removing duplicates, correcting errors, handling missing values, and standardizing data formats. For imaging data, preprocessing might involve resizing and augmenting images to improve model robustness. Textual data undergoes tokenization and the removal of irrelevant information to streamline analysis. Effective preprocessing ensures that the data fed into the AI model is clean, consistent, and ready for feature extraction [203].
The core of the development process is model training, where the AI model learns patterns and relationships within the preprocessed data [204]. The dataset is typically split into training, validation, and test sets to facilitate this learning. During training, the model iteratively adjusts its parameters to minimize errors. This phase is critical for building a model that generalizes well to new, unseen data. Following training, the model undergoes validation to fine-tune hyperparameters and prevent overfitting, ensuring it performs well not just on the training data but also on the validation set. The final step involves testing the model on the test set to evaluate its generalization capability and performance.
Importance of data quality and labeling
High-quality data and accurate labeling are paramount in AI model development. Data quality impacts the model's ability to learn meaningful patterns and make reliable predictions [205]. Poor-quality data, filled with errors, noise, or inconsistencies, can lead to inaccurate models that fail in real-world applications. In the context of food systems, where decisions can directly affect public health, maintaining high data quality is essential to ensure the model's outputs are trustworthy.
Data labeling is equally important, especially for supervised learning models [206]. Correctly labeled data allows the model to learn the correct associations between input features and the target output. For instance, in food quality control, labels might indicate whether a product is defective or safe. Inaccurate labels can mislead the model, resulting in poor performance and potentially dangerous misclassifications. Effective labeling often involves domain experts to ensure labels are precise and reflect real-world conditions accurately.
Techniques for model validation and testing
Model validation and testing [207] are critical steps to ensure an AI model's robustness and reliability. Cross-validation is a commonly used technique where the dataset is divided into multiple subsets, and the model is trained and validated on different combinations of these subsets. This helps ensure that the model's performance is consistent across various data segments and not reliant on a particular data split.
Hyperparameter tuning [208] is another essential aspect of validation, involving the adjustment of model parameters to achieve the best performance. Techniques such as grid search or random search can be used to systematically explore different hyperparameter values. Ensuring the model is not overfitting to the training data, techniques like regularization, dropout (for neural networks), and data augmentation are employed. These techniques help the model generalize better to unseen data.
Model testing [209] involves evaluating the final model on a separate test set, which was not used during training or validation. This provides an unbiased assessment of the model's performance. Metrics such as accuracy, precision, recall, F1 score, and mean squared error (depending on the task) are used to quantify the model's effectiveness. This step ensures the model's predictions are reliable and it can perform well in real-world applications.
In summary, developing effective AI models for food safety involves meticulous data collection, rigorous preprocessing, careful training and validation, and thorough testing. High data quality and accurate labeling are critical to model success, while robust validation and testing techniques ensure the model's reliability and generalizability in real-world scenarios.
Integration of AI with other technologies
IoT
Smart sensors and devices play a crucial role in enhancing food systems through real-time monitoring, providing continuous oversight of environmental conditions, processing parameters, and product quality. These technologies integrate seamlessly with AI systems to detect anomalies, predict potential issues, and ensure compliance with safety standards.
Environmental monitoring sensors: Smart sensors are deployed in food production and storage environments to monitor critical factors such as temperature, humidity, and gas levels [210, 211]. For instance, temperature sensors in refrigerated storage units ensure that perishable items are kept at optimal temperatures to prevent spoilage. Humidity sensors help maintain the appropriate moisture levels to avoid mold growth in storage facilities. By continuously collecting and transmitting data, these sensors enable real-time monitoring and immediate alerts if conditions deviate from preset thresholds.
Contamination detection sensors: Specialized sensors are used to detect contaminants such as pathogens, chemicals, and foreign objects in food products [212, 213]. For example, biosensors can identify microbial contamination by detecting specific bacteria or viruses. Chemical sensors can monitor for residues of pesticides or other hazardous substances. These sensors provide rapid detection, allowing for immediate corrective actions to be taken, thereby minimizing the risk of contaminated products reaching consumers.
Quality control devices: Smart devices equipped with advanced imaging and spectroscopy technologies are used for real-time quality assessment of food products [214, 215]. Hyperspectral imaging sensors, for example, can analyze the surface and internal composition of food items to detect defects, ripeness, and freshness. These devices can be integrated into production lines to automatically sort and remove defective products, ensuring only high-quality items proceed through the supply chain.
RFID and NFC Tags: Radio-frequency identification (RFID) [216] and Near Field Communication (NFC) tags [217] are used for tracking and tracing food products throughout the supply chain. These tags store information about the product’s origin, processing history, and storage conditions. Smart devices can read these tags to verify product authenticity, traceability, and compliance with safety standards. This real-time data collection and tracking enhance transparency and accountability in the food supply chain.
IoT-enabled smart packaging: Innovative smart packaging solutions integrate sensors and indicators directly into food packaging [218]. These smart packages can monitor the freshness and safety of the contents, providing visual indicators (such as color changes) if the product is compromised. For example, time–temperature indicators (TTIs) can show if a product has been exposed to temperatures outside its safe range during transportation or storage. This real-time feedback helps consumers and retailers make informed decisions about product safety.
Smart devices for process monitoring: In food processing plants, smart devices are used to monitor and control various production parameters, such as cooking temperatures, mixing speeds, and ingredient proportions [219]. These devices ensure that processes are carried out consistently and within safety guidelines. Real-time data from these devices is fed into AI systems, which can optimize processes, predict maintenance needs, and enhance overall efficiency.
AI-enhanced advanced food processing
Advanced food processing methods like ultrasonic sterilization, irradiation, and pulsed light treatment are increasingly incorporating AI technologies to ensure efficiency and safety. For example:
Hyperspectral imaging combined with AI detects product defects and microbial contamination at a microscopic level, enhancing inspection accuracy.
Robotics equipped with AI systems automate quality control for advanced processing lines, ensuring consistency and precision.
Predictive maintenance, driven by RL models, ensures continuous operation of high-tech processing equipment, preventing downtime and contamination risks.
These AI-integrated approaches not only improve food safety but also reduce energy use and enhance sustainability in modern food systems.
Blockchain technologies
Blockchain technologies are increasingly being adopted in the food industry to enhance transparency, traceability, and trust throughout the food supply chain. By providing a secure and immutable ledger of transactions, blockchain addresses several challenges in food safety and quality control. The different impacts of integrating blockchain technologies into the realm of food safety can be stated as follows:
Enhanced traceability: Blockchain enables end-to-end traceability by recording every transaction and movement of food products on a distributed ledger [220]. Each entry includes detailed information about the product’s origin, processing history, and handling conditions. This transparency allows stakeholders to trace a product’s journey from farm to table, quickly identifying the source of contamination or quality issues in the event of a recall. For instance, if a batch of lettuce is found to be contaminated, blockchain records can pinpoint the exact farm, processing facility, and distribution path, facilitating a targeted and efficient recall.
Improved food fraud prevention: Food fraud, such as counterfeiting and mislabeling, is a significant concern in the food industry. Blockchain provides a tamper-proof record of transactions and product information, making it difficult for fraudsters to alter or falsify data [221]. For example, blockchain can verify the authenticity of high-value products like olive oil or honey by ensuring that the product’s origin and processing details match the claims made on the label. Consumers and retailers can use blockchain to verify the integrity of products, reducing the risk of fraud.
Real-time monitoring and compliance: Blockchain can be integrated with IoT devices to monitor environmental conditions such as temperature and humidity during transportation and storage [222]. Data from these sensors is recorded on the blockchain, providing a real-time, immutable record of conditions throughout the supply chain. This continuous monitoring helps ensure that products are stored and transported under optimal conditions, maintaining their quality and safety. It also aids in regulatory compliance by providing a transparent record of adherence to safety standards.
Streamlined supply chain management: By providing a single and shared ledger accessible to all parties in the supply chain, blockchain simplifies coordination and communication [223]. Suppliers, processors, distributors, and retailers can access up-to-date information about product status, reducing delays and inefficiencies. For instance, blockchain can automate the verification of certifications and compliance documents, ensuring that all parties meet the required standards without the need for manual checks.
Consumer confidence and transparency: Blockchain enhances consumer trust by providing access to detailed information about food products [224]. Through blockchain-enabled platforms and applications, consumers can scan QR codes or access digital records to learn about a product’s origin, production methods, and safety history. This transparency empowers consumers to make informed choices and fosters confidence in the safety and quality of the food they purchase.
Efficient recall management: In the event of a food safety issue or recall, blockchain facilitates quick and efficient management. The detailed traceability provided by blockchain allows companies to identify affected products swiftly and accurately [225]. This efficiency reduces the scope and impact of recalls, minimizing waste and protecting public health. For example, if a batch of meat is found to be contaminated, blockchain records enable quick identification of all products affected and their distribution points.
Collaboration and data sharing: Blockchain fosters collaboration between different stakeholders in the food supply chain by providing a transparent and secure platform for data sharing [226]. This collaboration improves overall food safety by enabling stakeholders to share information about risks, best practices, and compliance efforts. It also facilitates data integration with other technologies, such as AI and IoT, to enhance food safety management.
Smart devices and robotics
Smart devices and robotics are transforming food security by introducing advanced technologies that enhance monitoring, efficiency, and quality control throughout the food supply chain. Smart devices are increasingly integrated into advanced food processing systems. For instance, robotic arms equipped with AI and hyperspectral imaging are used to detect microbial contamination during ultrasonic sterilization, ensuring safety and uniformity across batches. Additionally, automated systems optimize irradiation and pulsed electric field treatments by dynamically adjusting operational parameters based on AI-driven analysis of real-time data. By integrating IoT sensors, automated systems, and sophisticated robotics, these technologies provide real-time oversight, precise operations, and comprehensive tracking. This integration not only improves the safety and quality of food products but also optimizes resource management and supports effective emergency responses. Here are five major contributions of smart devices and robotics to food security:
Enhanced monitoring and control: Smart devices, including sensors and IoT technologies, provide real-time monitoring of critical factors such as temperature, humidity, and environmental conditions throughout the food supply chain [227]. This continuous oversight helps maintain optimal conditions, prevent spoilage, and ensure food safety. Automated alerts and data logging further enhance the ability to respond promptly to deviations and manage conditions effectively.
Automated quality assurance: Robotics and smart devices equipped with advanced imaging and sensor technologies automate quality control processes [228]. Robots perform precise inspections and sorting tasks, detecting defects, contaminants, and inconsistencies with high accuracy. This automation reduces human error, improves consistency, and ensures that only high-quality products reach consumers.
Improved traceability and tracking: IoT-enabled smart devices, such as RFID tags and QR codes, facilitate detailed tracking and traceability of food products throughout the supply chain [229]. This technology enhances transparency, allowing stakeholders to trace products from origin to point of sale. It also supports quick and efficient recall management in case of safety issues.
Efficient handling and processing: Robotics technology automates repetitive and labor-intensive tasks in food processing and packaging. Automated systems handle sorting, packing, and labeling with speed and precision, increasing efficiency and reducing contamination risks. In agriculture, robotic systems contribute to precision farming by optimizing planting, watering, and harvesting processes.
Enhanced resource management and emergency response: Smart devices and robotics support better resource management through precision agriculture and climate control systems [230]. Robotics also play a role in emergency response scenarios, such as delivering food supplies during crises or natural disasters. These technologies ensure timely assistance and efficient resource utilization, contributing to overall food security.
Case studies and real-world implementations
Several companies and organizations are utilizing AI to revolutionize various aspects of food systems, focusing on safety, quality, and security. These advancements optimize operations, ensure product integrity, and enhance traceability across the food supply chain. Here are some notable examples:
IBM: IBM's Watson is applied in food systems, particularly for traceability and supply chain management. IBM Food Trust leverages both AI and blockchain technology to enhance transparency and traceability, improving food safety and reducing the risks of contamination across the food system [231].
Microsoft: Through its Azure platform, Microsoft uses AI and ML to tackle food safety challenges. Their AI solutions facilitate real-time monitoring, predictive maintenance, and data analytics, which help food manufacturers and retailers ensure safety, maintain product quality, and optimize food system operations [232].
Nestlé: Nestlé employs AI to enhance food safety and quality through predictive analytics and automated quality control. By analyzing sensor data and imaging systems with AI algorithms, Nestlé improves the detection of contaminants and defects in food products, ensuring safety and quality across the food supply chain [233].
Cargill: Cargill applies AI to monitor and optimize various aspects of its food processing operations. Their AI-driven systems are integral in quality control, risk assessment, and predictive maintenance, ensuring that food products meet high safety and quality standards throughout the food system [234].
Zest Labs: Zest Labs focuses on managing the fresh food supply chain using AI and IoT technologies. Their Zest Fresh platform provides real-time monitoring and analytics to track food freshness and quality, which helps reduce spoilage and ensures compliance with safety standards across the entire supply chain [235].
Tetra Pak: Tetra Pak integrates AI within its food processing and packaging systems to strengthen quality control and ensure food safety. AI algorithms monitor and optimize processing conditions, detect anomalies, and safeguard the integrity of packaged products from production to consumer delivery [236].
Ecolab: Ecolab utilizes AI and data analytics to improve hygiene practices and food safety in food processing and handling environments. Their AI-driven solutions help identify potential risks, optimize cleaning protocols, and ensure overall food system safety measures [237].
Ripe Robotics: Ripe Robotics is developing AI-powered robots for automated harvesting and quality control within agriculture. These robots use CV and ML to identify and harvest ripe fruits, improving produce quality and reducing reliance on human labor, while optimizing agricultural food systems [238].
Clear Labs: Clear Labs applies AI and next-generation sequencing technologies to strengthen food safety testing. Their platform offers comprehensive analysis of food samples to detect contaminants and ensure compliance with food safety regulations, improving security throughout the food supply chain [239].
These companies and organizations demonstrate the diverse applications of AI in food safety, from enhancing traceability and quality control to optimizing processing and reducing contamination risks.
Challenges and limitations
Technical challenges
As illustrated in Fig. 5, AI implementation in food systems faces several challenges, particularly related to data availability and quality, high implementation costs, ethical concerns, and regulatory barriers. These challenges create obstacles to the widespread adoption of AI technologies aimed at enhancing food safety, quality, and security.
Fig. 5 [Images not available. See PDF.]
Challenges and opportunities in AI implementation for food safety, quality and security
Data quality: Data quality is a critical issue in food safety as it directly impacts the performance and accuracy of AI and ML models. Inconsistent, incomplete, or inaccurate data can lead to unreliable predictions and insights, potentially compromising food safety. For example, missing or erroneous sensor data can affect real-time monitoring and control systems, leading to undetected temperature deviations.
Algorithm limitations: AI algorithms, including those used in predictive analytics, image recognition, and anomaly detection, have inherent limitations [240]. For instance, ML models may struggle to generalize from training data to real-world scenarios, leading to reduced performance in detecting new or unforeseen types of defects and contaminants. Additionally, some algorithms may require extensive computational resources, which can be a limitation for organizations with limited technical infrastructure.
Integration with existing systems: Integrating AI technologies with existing food systems and processes can be complex [241]. Legacy systems may not be compatible with new AI solutions, leading to integration challenges. Furthermore, ensuring that AI systems work seamlessly with existing hardware, software, and workflows requires careful planning and implementation.
Algorithm transparency and interpretability: Many AI algorithms, particularly DL models, are often considered "black boxes," meaning their decision-making processes are not easily interpretable. This lack of transparency can be problematic in food safety, where understanding the rationale behind AI decisions is crucial for trust and compliance [242].
Scalability and adaptability: Scaling AI solutions to handle large volumes of data and adapt to changing conditions can be challenging [243]. As food safety needs evolve and data volumes increase, AI systems must be capable of scaling up and adapting to new types of data and scenarios.
Regulatory compliance and data privacy: Ensuring that AI systems comply with food safety regulations and data privacy laws is a significant challenge. Compliance with regulations such as the FDA’s Food Safety Modernization Act (FSMA) and data protection laws requires careful consideration of AI system design and data handling practices [244].
Ethical considerations
Privacy concerns
Issue: Privacy concerns arise when collecting and processing personal data, such as consumer feedback or health-related information, in food safety applications. The use of AI and data analytics often involves gathering extensive data on individuals, which can include sensitive information. Mismanagement of this data or unauthorized access can lead to privacy violations and loss of consumer trust [245].
Considerations: To address privacy concerns, organizations must adhere to data protection regulations such as the General Data Protection Regulation (GDPR) or the California Consumer Privacy Act (CCPA). Implementing strong data anonymization and encryption techniques can help protect personal information. Additionally, obtaining explicit consent from individuals before collecting their data and providing transparency about data usage are critical steps in maintaining privacy.
Data security
Issue: Ensuring data security is crucial to prevent unauthorized access, data breaches, and cyberattacks. Food safety systems often store and process large volumes of sensitive data, including supply chain information and quality control metrics. A breach or loss of this data can compromise food safety, disrupt operations, and damage an organization’s reputation.
Considerations: Robust cybersecurity measures, such as encryption, multi-factor authentication, and regular security audits, are essential to protect data. Organizations should also implement access controls to limit data access to authorized personnel only. Regularly updating software and systems to address vulnerabilities and staying informed about emerging threats can further enhance data security.
Transparency and accountability
Issue: Transparency and accountability in AI decision-making processes are vital to ensure ethical practices in food safety. AI systems can sometimes make decisions that are difficult to interpret or understand, raising concerns about the accountability of these decisions, especially when they impact consumer health and safety [246].
Considerations: Adopting explainable AI (XAI) methods can improve transparency by providing insights into how AI models make decisions. Clear documentation and communication about AI system processes and limitations are also important. Establishing accountability mechanisms, such as regular audits and review processes, helps ensure that AI systems are used responsibly and ethically.
Bias and fairness
Issue: Bias in AI systems can lead to unfair or discriminatory outcomes. For example, biased data used for training AI models can result in inaccurate predictions or quality assessments, potentially affecting certain groups unfairly. In food safety, bias can impact decision-making related to product quality, supplier evaluations, and compliance checks [247].
Considerations: To mitigate bias, it is important to use diverse and representative data sets when training AI models. Regularly evaluating and auditing AI systems for potential biases can help identify and address issues early. Incorporating fairness and inclusivity principles into AI system design and development is essential for achieving equitable outcomes.
Informed consent
Issue: Obtaining informed consent from individuals before collecting and using their data is a fundamental ethical consideration. In food safety applications, this involves ensuring that consumers and stakeholders are aware of how their data will be used and have the opportunity to opt-in or opt-out [248].
Considerations: Organizations should provide clear and concise information about data collection practices, purposes, and potential impacts. Obtaining explicit consent through straightforward and transparent consent forms is crucial. Ensuring that individuals have the option to withdraw consent at any time helps uphold ethical standards and respects individual autonomy.
Environmental impact
Issue: The deployment of AI and data collection technologies can have environmental implications, such as increased energy consumption and electronic waste. In the context of food safety, the use of sophisticated sensors and computing resources may contribute to the overall environmental footprint [249].
Considerations: To address environmental concerns, organizations should prioritize energy-efficient technologies and sustainable practices. Implementing strategies to reduce electronic waste, such as recycling and proper disposal of outdated equipment, can help minimize environmental impact. Considering the environmental impact in decision-making processes is essential for promoting sustainability.
Economic and practical constraints
Cost of Technology implementation
Constraint: Implementing advanced technologies, such as AI, IoT sensors, and robotics, can be prohibitively expensive. The initial investment for acquiring, integrating, and maintaining these systems can be significant, especially for small and medium-sized enterprises (SMEs). This includes costs related to hardware, software, training, and ongoing operational expenses [250].
Considerations: To manage costs, organizations can explore incremental implementation strategies, starting with pilot projects or scalable solutions. Leveraging cloud-based services and seeking government grants or industry subsidies can also help offset initial expenses. Collaboration with technology providers and industry partners may offer opportunities for cost-sharing and resource optimization.
Return on investment (ROI) uncertainty
Constraint: The uncertainty around the return on investment (ROI) from deploying advanced food safety technologies can be a significant barrier. Organizations may struggle to justify the expenditure if the expected benefits, such as improved safety or reduced spoilage, are not immediately clear or quantifiable [251].
Considerations: Conducting thorough cost–benefit analyses and pilot studies can help estimate the potential ROI and demonstrate the value of technology investments. Building a business case that highlights both tangible and intangible benefits, such as enhanced consumer trust and reduced regulatory risks, can support decision-making and justify expenditures.
Integration with legacy systems
Constraint: Integrating new technologies with existing legacy systems can be complex and costly. Many food safety operations rely on outdated systems that may not be compatible with modern technologies, leading to additional costs for integration, data migration, and process adjustments.
Considerations: To address integration challenges, organizations should conduct a comprehensive assessment of existing systems and develop a clear integration plan. Phased implementation and the use of middleware or APIs to bridge compatibility gaps can ease the transition. Collaborating with technology providers to ensure compatibility and minimize disruption is also essential.
Workforce training and skill development
Constraint: The adoption of new technologies often requires significant investment in workforce training and skill development. Employees need to be trained on new systems, processes, and technologies, which can be time-consuming and costly.
Considerations: Investing in targeted training programs and resources can help ensure that employees are equipped with the necessary skills. Developing comprehensive training materials and offering ongoing support can facilitate smoother transitions. Partnering with educational institutions or technology providers for specialized training can also be beneficial.
Regulatory compliance costs
Constraint: Ensuring compliance with food safety regulations and standards can incur additional costs. This includes expenses related to audits, certifications, and maintaining up-to-date compliance with evolving regulations. Non-compliance can also lead to fines and legal repercussions.
Scalability and adaptability issues
Constraint: Scaling new technologies to accommodate growing operations or adapting them to changing conditions can present challenges. As organizations expand or evolve, they may face difficulties in scaling up technology solutions or adjusting them to meet new requirements.
Considerations: Choosing scalable and flexible technologies that can grow with the organization is essential. Designing systems with modular components and adopting cloud-based solutions can facilitate scalability and adaptability. Regularly reviewing and updating technology strategies to align with organizational growth and changes can also help address scalability issues.
Data management and security
Constraint: Managing large volumes of data generated by advanced technologies poses economic and practical challenges. Ensuring data security and integrity requires investment in storage solutions, cybersecurity measures, and data management systems [252].
Considerations: Implementing efficient data management practices, such as data archiving and backup solutions, can help address data-related challenges. Investing in robust cybersecurity measures and leveraging cloud storage options can also enhance data security and manage costs effectively. Regular data audits and maintenance can ensure data quality and compliance.
Future directions and opportunities
Emerging trends in AI for food safety
Advanced predictive analytics
Trend: Predictive analytics, powered by AI, is increasingly being used to forecast potential food safety issues before they occur. By analyzing historical data, environmental conditions, and real-time monitoring inputs, AI models can predict risks such as contamination outbreaks, equipment failures, and spoilage events.
Implications: This trend allows for proactive measures to be taken, improving response times and reducing the impact of potential food safety issues. It also helps in optimizing resource allocation and enhancing overall operational efficiency.
Integration of AI with IoT and smart sensors
Trend: The integration of AI with IoT devices and smart sensors is becoming more prevalent. AI algorithms process data collected from various sensors embedded in food processing and storage environments to monitor conditions like temperature, humidity, and contamination levels in real-time.
Implications: This integration provides more accurate and timely insights into food safety conditions, leading to better monitoring and control. It enables dynamic adjustments to processes based on real-time data, improving overall food safety and quality.
Enhanced food traceability through Blockchain and AI
Trend: Combining AI with blockchain technology is enhancing food traceability and transparency in the supply chain. Blockchain provides an immutable record of transactions, while AI analyzes and verifies the data to ensure authenticity and traceability of food products from farm to table.
Implications: This trend improves consumer trust and confidence by providing verifiable information about the origins and handling of food products. It also facilitates quicker response to recalls and enhances compliance with regulatory requirements.
Development of Explainable AI (XAI)
Trend: XAI is gaining traction to address the "black box" issue of traditional AI models. XAI focuses on making AI decision-making processes more transparent and understandable to users by providing clear explanations of how decisions are made [253].
Implications: This trend enhances trust and accountability in AI systems used for food safety. It allows stakeholders to understand and verify AI-driven decisions, which is crucial for compliance and regulatory reporting.
AI-Driven personalized nutrition
Trend: AI is being used to develop personalized nutrition solutions based on individual dietary needs and health conditions [254]. By analyzing data from various sources, including genetic information, lifestyle factors, and dietary habits, AI can provide tailored recommendations to improve individual health and prevent foodborne illnesses.
Implications: Personalized nutrition enhances food safety by ensuring that individuals receive diet recommendations suited to their specific needs, potentially reducing the risk of adverse health outcomes related to food consumption.
Advanced image recognition and CV
Trend: AI-driven image recognition and CV technologies are being increasingly applied to food systems for tasks such as quality control, defect detection, and contamination identification. These technologies analyze visual data from food products and processing environments to detect anomalies and ensure adherence to safety standards.
Implications: Improved accuracy in detecting defects and contaminants enhances overall food safety and quality control. It also reduces the reliance on manual inspection, leading to more efficient and consistent quality assessments.
AI for food fraud detection
Trend: AI is being employed to combat food fraud by analyzing data related to food products, supply chains, and consumer behavior. ML algorithms detect anomalies and patterns indicative of fraudulent activities, such as mislabeling or adulteration.
Implications: This trend helps protect consumers from food fraud and ensures the authenticity and integrity of food products. It also supports regulatory compliance and strengthens food safety measures.
Robotics and automation in food safety
Trend: The use of robotics and automation, powered by AI, is expanding in food processing and handling. Automated systems equipped with AI are used for tasks such as sorting, packing, and inspection, improving efficiency and consistency in food safety practices.
Implications: Robotics and automation enhance operational efficiency, reduce human error, and maintain high standards of food safety. They also address labor shortages and improve scalability in food production and processing.
AI for environmental monitoring
Trend: AI technologies are increasingly being used for environmental monitoring in food safety. AI models analyze data from environmental sensors to track conditions such as air quality, temperature, and humidity, which are critical for maintaining food safety in production and storage environments.
Implications: Real-time environmental monitoring ensures optimal conditions for food safety, reduces spoilage, and prevents contamination. It also supports compliance with environmental regulations and standards.
Use of AI in regulatory compliance and documentation
Trend: AI is being used to streamline regulatory compliance and documentation processes. AI systems automate the analysis of regulatory texts, compliance records, and inspection reports to ensure adherence to food safety regulations and standards.
Implications: This trend improves efficiency in regulatory compliance, reduces manual effort, and enhances accuracy in tracking and reporting. It also supports organizations in staying up-to-date with evolving regulations and standards.
Research gaps and areas for future study
Data quality and standardization
Gap: Despite significant advancements in AI, inconsistencies in data quality and the lack of standardized formats across the food system remain major challenges. Variability in data collection techniques, sensor calibration, and data reporting standards can undermine the reliability of AI models used in food systems.
Future study: Research should focus on developing standardized protocols for data collection, integration, and management across all levels of the food system. Additionally, advancing techniques for data cleaning, validation, and harmonization can significantly enhance the performance and reliability of AI models for food safety, quality, and security.
Explainability and transparency
Gap: Many AI models, particularly those based on DL, function as "black boxes," making it difficult to understand their decision-making processes. This opacity can reduce trust and hinder the adoption of AI in broader food systems, including food safety and quality control.
Future Study: Investigating methods for enhancing the explainability and transparency of AI models is crucial. Research should focus on developing XAI techniques that provide clear insights into the decision-making process of AI models, which can increase trust among stakeholders and ease regulatory compliance across the food system.
Integration with existing systems
Gap: Integrating AI technologies with legacy systems in food safety can be complex and costly. Many existing systems are not designed to interface with modern AI solutions, leading to challenges in seamless integration.
Future study: Research should explore innovative integration techniques, such as middleware solutions and modular system designs, that facilitate compatibility between AI technologies and existing infrastructure. Additionally, developing cost-effective and scalable integration strategies can support wider adoption.
Bias and fairness
Gap: AI models can inadvertently introduce or perpetuate biases, especially if training data is not diverse or representative. This can lead to unfair outcomes and impact the effectiveness of food safety measures.
Future study: Future research should focus on identifying and mitigating sources of bias in AI models. Developing methods for ensuring fairness and inclusivity in training data and model outcomes is essential for equitable food safety solutions.
Real-time data processing and scalability
Gap: Processing large volumes of real-time data from various sources, such as IoT sensors and imaging systems, presents scalability challenges. Ensuring timely and accurate analysis of data remains a significant obstacle.
Future study: Research into scalable data processing architectures and efficient algorithms for real-time analysis is needed. Exploring edge computing solutions and distributed processing frameworks can improve the handling of large-scale data and enhance system performance.
Privacy and security concerns
Gap: The collection and analysis of sensitive data, such as consumer health information and supply chain details, raise privacy and security concerns. Ensuring data protection while leveraging AI for food safety is a complex challenge.
Future study: Investigating advanced privacy-preserving techniques, such as differential privacy and secure multi-party computation, can help address security concerns. Research should also focus on developing robust cybersecurity measures to protect data integrity and confidentiality.
Interdisciplinary collaboration
Gap: Effective implementation of AI in food safety often requires collaboration across various disciplines, including data science, food science, and regulatory affairs [255]. Lack of interdisciplinary communication can hinder the development of comprehensive solutions.
Future study: Promoting interdisciplinary research and collaboration between AI experts, food safety professionals, and regulatory bodies is essential. Developing frameworks for effective communication and cooperation can facilitate the integration of AI into food safety practices.
Ethical and social implications
Gap: The ethical and social implications of deploying AI in food safety, including issues related to data privacy, consent, and accountability, are not fully explored [256]. Understanding these implications is crucial for responsible AI adoption.
Future study: Research should address the ethical considerations associated with AI in food safety, including privacy, consent, and the social impact of automation. Developing guidelines and best practices for ethical AI use can support responsible implementation.
Long-term impact assessment
Gap: There is limited research on the long-term impacts of AI implementation in food safety, including its effects on food quality, safety outcomes, and industry practices.
Future study: Conducting longitudinal studies to assess the long-term impacts of AI technologies on food safety and industry practices is essential. Evaluating the sustained effectiveness and potential unintended consequences of AI interventions can provide valuable insights for future development.
Cost–benefit analysis
Gap: The economic implications of adopting AI technologies in food safety, including cost–benefit analysis and ROI, are not thoroughly examined. Understanding the financial impact of AI investments is crucial for decision-making.
Future study: Research should focus on comprehensive cost–benefit analyses of AI technologies in food safety. Evaluating the economic benefits, including cost savings, efficiency gains, and risk reduction, can support informed investment decisions and adoption strategies.
Conclusions
This review provides a comprehensive analysis of the application of AI in food systems, consolidating evidence on the use of ML, NLP, CV, and RL for improving risk detection, monitoring, and traceability across food supply chains. While AI technologies have been widely studied in different food systems, this review offers a deeper insight into the integration of AI methodologies with food processing techniques and the practical challenges that limit their broader adoption. The analysis reveals several key contributions and areas for innovation based on current literature and ongoing industry needs.
Main contributions:
Systematic integration of AI technologies: This review synthesizes the fragmented research on AI applications in food systems, highlighting how AI-driven systems can be integrated across multiple stages of the supply chain—from production to consumption—thereby addressing gaps in real-time hazard detection, traceability, and processing optimization.
AI in food valorization: A key highlight of this review is the application of AI in enhancing food valorization processes, such as converting food waste and by-products into high-value materials. AI-driven optimizations in these processes enhance efficiency, reduce waste, and promote sustainability in modern food systems.
Practical application insights: Unlike previous studies that focus on isolated AI technologies, this review evaluates the synergy between AI and other emerging technologies, such as food processing techniques. AI's role in optimizing methods like high-pressure processing, UV treatment, and pulsed light treatment is explored, emphasizing its potential to enhance processing efficiency and food safety while minimizing risks.
Risk mitigation in dynamic environments: The review highlights AI’s ability to adapt to complex and dynamic environments where contamination risks evolve over time. This adaptability offers unique advantages over traditional methods, especially in optimizing food processing conditions, identifying contamination sources, and improving predictive modeling for emerging pathogens.
Limitations of AI models:
While AI presents significant potential, the review highlights several limitations that need to be addressed to ensure its effectiveness in food systems:
Data quality and availability: A significant barrier to successful AI implementation is the inaccessibility of high-quality, comprehensive datasets required for training robust models. Current AI models often rely on isolated or incomplete data, which limits their accuracy and predictive capabilities, particularly in complex food processing environments.
Model reliability under uncertainty: AI models struggle with uncertainty in environments where data may be missing, incomplete, or subject to noise. The reliability of AI systems in these uncertain contexts needs further exploration, especially in food processing, where such situations are common.
Ethical concerns and adoption costs: The review also highlights critical ethical concerns, including data privacy, bias in AI decision-making, and the lack of transparency in AI models. Additionally, high implementation costs create barriers, especially for small to medium-sized enterprises (SMEs), limiting AI’s democratization in the food industry.
Future research directions:
Building on the insights from the current literature, this review identifies several areas for future research and innovation that can enhance the utility and adoption of AI in food systems:
Development of global standardized datasets: To address the limitations of data quality and availability, future work should focus on creating open-source, standardized datasets that can be used to train more reliable AI models. Collaborative efforts between governments, industry stakeholders, and academia will be crucial in this endeavor.
Improving model robustness and transparency: AI models must be improved for reliability under uncertain conditions. Developing models that can handle incomplete or noisy data and still provide accurate predictions will significantly enhance their utility in real-world applications, particularly in food processing. Additionally, ensuring that these models are transparent and interpretable is critical to increasing trust and adoption.
Scaling AI technologies for wider adoption: Research should also focus on reducing the economic barriers to AI adoption. This can be achieved by developing cost-effective and scalable AI solutions that cater to the needs of smaller producers and resource-constrained environments. Leveraging cloud computing and distributed AI systems may offer scalable solutions that reduce costs.
Integration with emerging food processing technologies: Future research should explore how AI can be more deeply integrated with emerging food processing technologies such as high-pressure processing, UV treatment, and pulsed light treatment. AI-driven optimizations in these processes could lead to more resilient food safety infrastructures, enabling faster responses to contamination risks, enhancing quality control, and reducing processing time.
By addressing these limitations and focusing on the identified areas for future research, AI has the potential to revolutionize food safety management and food processing. With continued advancements, AI can deliver more efficient, scalable, and reliable food systems that proactively mitigate risks and enhance global food security.
Author contribution
S.B.D and D.K. wrote the manuscript, prepared figures and have reviewed the manuscript.
Funding
This research was funded by the Department of Analytical Chemistry under the Directorate of Energy, Environment, Science & Technology (EES&T) at Idaho National Laboratory (US Department of Energy).
Data availability
No datasets were generated or analysed during the current study.
Declarations
Competing interests
The authors declare no competing interests.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
1. Vågsholm, I; Arzoomand, NS; Boqvist, S. Food security, safety, and sustainability—getting the trade-offs right. Front Sustain Food Syst; 2020; 4, 16. [DOI: https://dx.doi.org/10.3389/fsufs.2020.00016]
2. Lin, C-TJ. Demographic and socioeconomic influences on the importance of food safety in food shopping. Agric Resour Econ Rev; 1995; 24, pp. 190-198.
3. Griffith, CJ. Do businesses get the food poisoning they deserve? The importance of food safety culture. Br Food J; 2010; 112, pp. 416-425.
4. Borchers, A et al. Food safety. Clin Rev Allergy Immunol; 2010; 39, pp. 95-141.
5. World Health Organization. The Work of WHO in the eastern Mediterranean region: annual report of the regional director, 1 January-31 December 1998; WHO: Geneva, Switzerland, 1999
6. Flynn, K et al. An introduction to current food safety needs. Trends Food Sci Technol; 2019; 84, pp. 1-3.
7. Oguntoyinbo, FA. Safety challenges associated with traditional foods of West Africa. Food Rev Int; 2014; 30, pp. 338-358. [DOI: https://dx.doi.org/10.1080/87559129.2014.940086]
8. Ashrafudoulla, M et al. Challenges and opportunities of non-conventional technologies concerning food safety. World’s Poult Sci J; 2023; 79, pp. 3-26. [DOI: https://dx.doi.org/10.1080/00439339.2023.2163044]
9. Tao, Y et al. High-pressure processing of foods: an overview. Emerging technologies for food processing; 2014; Amsterdam, Elsevier: pp. 3-24.
10. Delorme, MM et al. Ultraviolet radiation: an interesting technology to preserve quality and safety of milk and dairy foods. Trends Food Sci Technol; 2020; 102, pp. 146-154.
11. Bhavya, ML; Hebbar, HU. Pulsed light processing of foods for microbial safety. Food Qual Saf; 2017; 1, pp. 187-202.
12. Mason, TJ et al. Ultrasonic food processing. Alternatives to conventional food processing; 2011; Berlin, Springer: pp. 387-414.
13. Vashisht, P et al. Pulsed electric field processing in the dairy sector: a review of applications, quality impact and implementation challenges. Int J Food Sci Technol; 2024; 59, pp. 2122-2135.
14. Farkas, J; Mohácsi-Farkas, C. History and future of food irradiation. Trends Food Sci Technol; 2011; 22,
15. Sommers, C.H. and Fan, X. eds., 2008. Food irradiation research and technology, John Wiley & Sons.
16. Vashisht, P.; et al. Ozone Processing in the Dairy Sector: A Review of Applications, Quality Impact and Implementation Challenges. Food Sci. Technol. 2023.
17. Habib, MR et al. Alleviating heavy metal toxicity in milk and water through a synergistic approach of absorption technique and high voltage atmospheric cold plasma and probable rheological changes. Biomolecules; 2022; 12, 913. [DOI: https://dx.doi.org/10.3390/biom12070913]
18. Mahanta, S; Habib, MR; Moore, JM. Effect of high-voltage atmospheric cold plasma treatment on germination and heavy metal uptake by soybeans (Glycine max). Int J Mol Sci; 2022; 23, 1611.
19. Mahanta, S et al. Atmospheric cold plasma as an alternative to chlorination in soft wheat flour to prepare high-ratio cakes. Foods; 2024; 13, 2366.
20. Vashisht, P; Verma, D; Singh, L; Saini, GS; Sharma, S; Charles, AP; Mahanta, S; Mahanta, S; Singh, K; Gaurav, G; Kaur, J. Ozone processing of milk and milk products: a review of applications, quality effect and implementation challenges. Int J Food Eng; 2024; 20,
21. Havelaar, AH et al. Future challenges to microbial food safety. Int J Food Microbiol; 2010; 139, pp. S79-S94.
22. Bhat, R; Gómez-López, VM. Practical food safety: contemporary issues and future directions; 2014; Hoboken, Wiley:
23. Motarjemi, Y; Lelieveld, H. Fundamentals in management of food safety in the industrial setting: challenges and outlook of the 21st century. Food safety management; 2014; Cambridge, Academic Press: pp. 1-20.
24. Valdramidis, VP; Koutsoumanis, KP. Challenges and perspectives of advanced technologies in processing, distribution and storage for improving food safety. Curr Opin Food Sci; 2016; 12, pp. 63-69. [DOI: https://dx.doi.org/10.1016/j.cofs.2016.08.008]
25. Yadav, S et al. Valorisation of agri-food waste for bioactive compounds: recent trends and future sustainable challenges. Molecules; 2024; 29, 2055.
26. Kudashkina, K et al. Artificial intelligence technology in food safety: a behavioral approach. Trends Food Sci Technol; 2022; 123, pp. 376-381.
27. Liu, Z et al. Artificial intelligence in food safety: a decade review and bibliometric analysis. Foods; 2023; 12, 1242.
28. Friedlander, A; Zoellner, C. Artificial intelligence opportunities to improve food safety at retail. Food Prot Trends; 2020; 40, 4.
29. Qian, C et al. How can AI help improve food safety?. Annu Rev Food Sci Technol; 2023; 14, pp. 517-538. [DOI: https://dx.doi.org/10.1146/annurev-food-060721-013815]
30. Karanth, S et al. Importance of artificial intelligence in evaluating climate change and food safety risk. J Agric Food Res; 2023; 11,
31. Mu, W et al. Making food systems more resilient to food safety risks by including artificial intelligence, big data, and internet of things into food safety early warning and emerging risk identification tools. Compreh Rev Food Sci Food Saf; 2024; 23, [DOI: https://dx.doi.org/10.1111/1541-4337.13296] e13296.
32. Chhetri, KB. Applications of artificial intelligence and machine learning in food quality control and safety assessment. Food Eng Rev; 2024; 16, pp. 1-21.
33. Addanki, M; Patra, P; Kandra, P. Recent advances and applications of artificial intelligence and related technologies in the food industry. Appl Food Res; 2022; 2, [DOI: https://dx.doi.org/10.1016/j.afres.2022.100126] 100126.
34. Alpaydin, E. Machine learning; 2021; Cambridge, MIT Press:
35. Jordan, MI; Mitchell, TM. Machine learning: trends, perspectives, and prospects. Science; 2015; 349, pp. 255-260.
36. Wang, X et al. Application of machine learning to the monitoring and prediction of food safety: a review. Compreh Rev Food Sci Food Saf; 2022; 21, pp. 416-434.
37. Deng, X; Cao, S; Horn, AL. emerging applications of machine learning in food safety. Annu Rev Food Sci Technol; 2021; 12, pp. 513-538.
38. Said, Z et al. Intelligent approaches for sustainable management and valorisation of food waste. Biores Technol; 2023; 377,
39. Sabater, C et al. Vegetable waste and by-products to feed a healthy gut microbiota: current evidence, machine learning and computational tools to design novel microbiome-targeted foods. Trends Food Sci Technol; 2021; 118, pp. 399-417.
40. Nirmal, N et al. Enhancement in the active food packaging system through metal-based nanomaterials: a review of innovations, challenges, and future directions. Discov Food; 2024; 4,
41. LeCun, Y; Bengio, Y; Hinton, G. Deep learning. Nature; 2015; 521, pp. 436-444.
42. Goodfellow, I; Bengio, Y; Courville, A. Deep learning; 2016; Cambridge, MIT Press:
43. Kelleher, JD. Deep Learning; 2019; Cambridge, MIT Press:
44. Makridis, G.; et al. Enhanced food safety through deep learning for food recalls prediction. In: discovery science: 23rd international conference, DS 2020; Springer International Publishing: Cham, Switzerland, 2020; pp. 119–132.
45. Zhou, L et al. Application of deep learning in food: a review. Compreh Rev Food Sci Food Saf; 2019; 18, pp. 1793-1811.
46. Chen, T-C; Yu, S-Y. Research on food safety sampling inspection system based on deep learning. Food Sci Technol; 2021; 42, [DOI: https://dx.doi.org/10.1590/fst.29121] e29121.
47. Wang, Y.; et al. (2021) Food Image recognition and food safety detection method based on deep learning. Comput. Intell. Neurosci 1268453
48. Chowdhary, KR; Chowdhary, KR. Natural language processing in fundamentals of artificial intelligence; 2020; Cham, Springer: pp. 603-649.
49. Nadkarni, PM; Ohno-Machado, L; Chapman, WW. natural language processing: an introduction. J Am Med Inform Assoc; 2011; 18, pp. 544-551.
50. Eisenstein, J. Introduction to natural language processing; 2019; Cambridge, MIT Press:
51. Molenaar, A et al. Using natural language processing to explore social media opinions on food security: sentiment analysis and topic modeling study. J Med Internet Res; 2024; 26,
52. Martin, NM et al. Enhancing artificial intelligence for twitter-based public discourse on food security during the COVID-19 pandemic. Disaster Med Public Health Prep; 2022; 1, pp. 1-25. [DOI: https://dx.doi.org/10.1017/dmp.2022.207]
53. Benites-Lazaro, LL; Giatti, L; Giarolla, A. Topic modeling method for analyzing social actor discourses on climate change, energy and food security. Energy Res Soc Sci; 2018; 45, pp. 318-330. [DOI: https://dx.doi.org/10.1016/j.erss.2018.07.031]
54. Voulodimos, A et al. Deep learning for computer vision: a brief review. Comput Intell Neurosci; 2018; 2018, 7068349. [DOI: https://dx.doi.org/10.1155/2018/7068349]
55. Szeliski, R. Computer vision: algorithms and applications; 2022; Cham, Switzerland, Springer Nature:
56. Granlund, GH; Knutsson, H. Signal processing for computer vision; 2013; Berlin, Springer Science & Business Media:
57. Sood, S; Singh, H. Computer vision and machine learning based approaches for food security: a review. Multimedia Tools Appl; 2021; 80, pp. 27973-27999.
58. Kakani, V et al. A critical review on computer vision and artificial intelligence in food industry. J Agric Food Res; 2020; 2,
59. Fracarolli, JA et al. Computer vision applied to food and agricultural products. Rev Ciência Agronômica; 2020; 51, e20207749.
60. Squara, S. et al. Valorisation of premium Italian wines by volatile signature exploration with GC× GC-TOF MS and Computer Vision. In: 14th Multidimensional Chromatography Workshop Guide Book, 2023 (pp. 74–74).
61. Russom, P. Big data analytics. TDWI Best Pract Rep; 2011; 19, pp. 1-34.
62. Marvin, HJP et al. Big data in food safety: an overview. Crit Rev Food Sci Nutr; 2017; 57, pp. 2286-2295.
63. Jin, C et al. Big data in food safety—a review. Curr Opin Food Sci; 2020; 36, pp. 24-32.
64. Donaghy, JA et al. Big data impacting dynamic food safety risk management in the food chain. Front Microbiol; 2021; 12, [DOI: https://dx.doi.org/10.3389/fmicb.2021.668196] 668196.
65. Sharma, S et al. Sustainable innovations in the food industry through artificial intelligence and big data analytics. Logistics; 2021; 5, 66.
66. Rose, K; Eldridge, S; Chapin, L. The internet of things: an overview. Int Soc (ISOC); 2015; 80, pp. 1-53.
67. Li, S; Xu, LD; Zhao, S. The internet of things: a survey. Inf Syst Front; 2015; 17, pp. 243-259.
68. Xia, F et al. Internet of things. Int J Commun Syst; 2012; 25, pp. 1101-1111.
69. Bouzembrak, Y et al. Internet of things in food safety: literature review and a bibliometric analysis. Trends Food Sci Technol; 2019; 94, pp. 54-64.
70. Liu, Y et al. An Internet-of-things solution for food safety and quality control: a pilot project in China. J Ind Inf Integr; 2016; 3, pp. 1-7.
71. Doinea, M et al. Internet of things based systems for food safety management. Informatica Econ; 2015; 19, pp. 87-94. [DOI: https://dx.doi.org/10.12948/issn14531305/19.1.2015.08]
72. Dias, R.M.; Marques, G.; Bhoi, A.K. Internet of things for enhanced food safety and quality assurance: a literature review. In: proceedings of the international conference on emerging trends and advances in electrical engineering and renewable energy. Springer, Singapore, 2020; pp. 1–8.
73. Sharma, A et al. Recent trends in AI-based intelligent sensing. Electronics; 2022; 11, 1661.
74. Palakurti, NR. AI applications in food safety and quality control. ESP J Eng Technol Adv; 2022; 2, pp. 48-61.
75. Ataei Kachouei, M; Kaushik, A; Ali, MA. Internet of things-enabled food and plant sensors to empower sustainability. Adv Intell Syst; 2023; 5, 2300321. [DOI: https://dx.doi.org/10.1002/aisy.202300321]
76. Taneja, A et al. Artificial intelligence: implications for the agri-food sector. Agronomy; 2023; 13, 1397.
77. Caldwell, DG. Robotics and automation in the food industry: current and future technologies; 2012; Amsterdam, Elsevier:
78. Raheem, D; Treiblmaier, H; Mohammed, WM; Ferrer, BR; Martinez-Lastra, JL. Robotics as key enabler technology in food industry 4.0 and beyond. Food industry 4.0; 2024; Elsevier: pp. 121-131. [DOI: https://dx.doi.org/10.1016/B978-0-443-15516-1.00007-4]
79. Prasad, S. Application of robotics in dairy and food industries: a review. Int J Sci Environ Technol; 2017; 6, pp. 1856-1864.
80. Mor, RS et al. Robotics and automation for agri-food 4.0: innovation and challenges. Agri-food 4.0: innovations challenges and strategies; 2022; Bingley, Emerald Publishing Limited: pp. 189-199.
81. Iqbal, J; Khan, ZH; Khalid, A. Prospects of robotics in food industry. Food Sci Technol Camp; 2017; 37, pp. 159-165.
82. Pech, M; Vrchota, J; Bednář, J. Predictive maintenance and intelligent sensors in smart factory. Sensors; 2021; 21, 1470.
83. Rejeb, A et al. Blockchain technology in the food industry: a review of potentials. Chall Future Res Dir Logist; 2020; 4, 27.
84. Olsen, P.; Borit, M.; Syed, S. Applications, limitations, costs, and benefits related to the use of blockchain technology in the food industry. Nofima Rapportserie 2019.
85. Fortuna, F; Risso, M. Blockchain technology in the food industry. Symphonya Emerg Issues Manag; 2019; 2, pp. 151-158.
86. Patelli, N; Mandrioli, M. Blockchain technology and traceability in the agrifood industry. J Food Sci; 2020; 85, pp. 3670-3678.
87. Antonucci, F et al. A review on blockchain applications in the agri-food sector. J Sci Food Agric; 2019; 99, pp. 6129-6138.
88. Shahbazi, Z; Byun, Y-C. A procedure for tracing supply chains for perishable food based on blockchain. Mach Learn Fuzzy Logic Electron; 2020; 10, 41.
89. Polonsky, JA et al. Outbreak analytics: a developing data science for informing the response to emerging pathogens. Philos Trans R Soc B; 2019; 374, 20180276.
90. Wang, H et al. Machine learning prediction of foodborne disease pathogens: algorithm development and validation study. JMIR Med Inform; 2021; 9, [DOI: https://dx.doi.org/10.2196/24924] e24924.
91. Van Boekel, MAJS. Kinetics of heat-induced changes in dairy products: developments in data analysis and modelling techniques. Int Dairy J; 2022; 126,
92. Zhu, L-Y et al. Evaluation of methods for the detection of hazardous substances in food based on machine learning. New J Chem; 2024; 48, pp. 1399-1406.
93. Astuti, SD et al. Gas sensor array to classify the chicken meat with E. Coli contaminant by using random forest and support vector machine. Biosens. Bioelectron X; 2021; 9, 100083.
94. Torres-Sánchez, R et al. Real-time monitoring system for shelf life estimation of fruit and vegetables. Sensors; 1860; 2020, 20.
95. Fu, B; Labuza, TP. Shelf-life prediction: theory and application. Food Control; 1993; 4, pp. 125-133.
96. Martinez-Castillo, C et al. Random forest, artificial neural network, and support vector machine models for honey classification. EFood; 2020; 1, pp. 69-76.
97. de Santana, FB; Borges Neto, W; Poppi, RJ. Random forest as one-class classifier and infrared spectroscopy for food adulteration detection. Food Chem; 2019; 293, pp. 323-332. [DOI: https://dx.doi.org/10.1016/j.foodchem.2019.04.073]
98. Saha, D; Manickavasagan, A. Machine learning techniques for analysis of hyperspectral images to determine quality of food products: a review. Curr Res Food Sci; 2021; 4, pp. 28-44.
99. Mihirsen, D.D.; Joseph, J.T.; Renisha, B. Time series analysis for supply chain planning in restaurants. In: proceedings of the 2020 5th international conference on computing, communication and security (ICCCS); IEEE: New York, NY, USA, 2020.
100. Dellino, G et al. A reliable decision support system for fresh food supply chain management. Int J Prod Res; 2018; 56, pp. 1458-1485.
101. Ebel, ED et al. Comparing characteristics of sporadic and outbreak-associated foodborne illnesses, united states, 2004–2011. Emerg Infect Dis; 2016; 22, pp. 1193-1201.
102. Heisterkamp, SH; Dekkers, ALM; Heijne, JCM. Automated detection of infectious disease outbreaks: hierarchical time series models. Stat Med; 2006; 25, pp. 4179-4196.
103. Bonah, E et al. Electronic nose classification and differentiation of bacterial foodborne pathogens based on support vector machine optimized with particle swarm optimization algorithm. J Food Process Eng; 2019; 42, [DOI: https://dx.doi.org/10.1111/jfpe.13236] e13236.
104. Agany, DM; Pietri, JE; Gnimpieba, EZ. Assessment of vector-host-pathogen relationships using data mining and machine learning. Comput Struct Biotechnol J; 2020; 18, pp. 1704-1721. [DOI: https://dx.doi.org/10.1016/j.csbj.2020.06.031]
105. Krishna, K; Murty, MN. Genetic K-means algorithm. IEEE Trans Syst Man Cybern B Cybern; 1999; 29, pp. 433-439.
106. Maćkiewicz, A; Ratajczak, W. Principal components analysis (PCA). Comput Geosci; 1993; 19, pp. 303-342. [DOI: https://dx.doi.org/10.1016/0098-3004(93)90090-R]
107. Waite, RD et al. Clustering of pseudomonas aeruginosa transcriptomes from planktonic cultures, developing and mature biofilms reveals distinct expression profiles. BMC Genomics; 2006; 7, pp. 1-14.
108. Cai, Y; Sun, Y. ESPRIT-tree: hierarchical clustering analysis of millions of 16S rRNA pyrosequences in quasilinear computational time. Nucleic Acids Res; 2011; 39, [DOI: https://dx.doi.org/10.1093/nar/gkr349] e95.
109. Hejazi, M; Singh, YP. One-class support vector machines approach to anomaly detection. Appl Artif Intell; 2013; 27, pp. 351-366.
110. Priyanto, C.Y.; Purnomo, H.D. Combination of isolation forest and LSTM autoencoder for anomaly detection. In: proceedings of the 2021 2nd international conference on innovative and creative information technology (ICITech); IEEE: New York, NY, USA, 2021.
111. Petkovski, A.; Shehu, V. Anomaly detection on univariate sensing time series data for smart aquaculture using K-means, isolation forest, and local outlier factor. In: proceedings of the 2023 12th Mediterranean conference on embedded computing (MECO); IEEE: New York, NY, USA, 2023.
112. Zhou, Y. Application of K-Means clustering algorithm in fresh food safety management. In: proceedings of the international conference on applications and techniques in cyber intelligence; Springer International Publishing: Cham, Switzerland, 2022.
113. Pampoukis, G et al. Recent advances and applications of rapid microbial assessment from a food safety perspective. Sensors; 2022; 22, 2800.
114. Hansen, T; Wilke, R; Zaichkowsky, J. Managing consumer complaints: differences and similarities among heterogeneous retailers. Int J Retail Distrib Manag; 2010; 38, pp. 6-23. [DOI: https://dx.doi.org/10.1108/09590551011016304]
115. Singh, A; Shukla, N; Mishra, N. Social media data analytics to improve supply chain management in food industries. Transp Res Part E Logist Transp Rev; 2018; 114, pp. 398-415.
116. Kaelbling, LP; Littman, ML; Moore, AW. Reinforcement learning: a survey. J Artif Intell Res; 1996; 4, pp. 237-285.
117. Wiering, MA; Van Otterlo, M. Reinforcement learning. Adaptation, learning, and optimization; 2012; Berlin/Heidelberg, Springer: pp. 729-775.
118. Wu, M; Liu, W; Zheng, S. Intelligent food safety: a prediction model based on attention mechanism and reinforcement learning. Appl Artif Intell; 2024; 38, 2379731. [DOI: https://dx.doi.org/10.1080/08839514.2024.2379731]
119. Barendsz, AW. Food safety and total quality management. Food Control; 1998; 9, pp. 163-170.
120. Lermen, F.H.; et al. (2023) reinforcement learning system to capture value from Brazilian post-harvest Offers. Inform. Process. Agric
121. Chen, H et al. Effective management for blockchain-based agri-food supply chains using deep reinforcement learning. IEEE Access; 2021; 9, pp. 36008-36018. [DOI: https://dx.doi.org/10.1109/ACCESS.2021.3062410]
122. Wang, X. The analysis and re-optimization of food systems by using intelligent optimization algorithms and machine learning. All Life; 2022; 15, pp. 656-677.
123. Allik, B.; Shedleski, F.; Hsu, C. Learning control actions for guided projectiles using proximal policy optimization (PPO) algorithm. ARL-TR-9603 2022.
124. Aboutorab, H et al. A reinforcement learning-based framework for disruption risk identification in supply chains. Future Gener Comput Syst; 2022; 126, pp. 110-122. [DOI: https://dx.doi.org/10.1016/j.future.2021.08.004]
125. Hsieh, P; Benros, D; Dogan, T. Conversational co-creativity with deep reinforcement learning agent in kitchen layout. Design computing and cognition’20; 2022; Cham, Springer International Publishing: pp. 399-409.
126. Li, Z et al. A survey of convolutional neural networks: analysis, applications, and prospects. IEEE Trans Neural Netw Learn Syst; 2021; 33, pp. 6999-7019.
127. Grossberg, S. Recurrent neural networks. Scholarpedia; 1888; 2013, 8.
128. Kollia, I; Stevenson, J; Kollias, S. AI-enabled efficient and safe food supply chain. Electronics; 2021; 10, 1223. [DOI: https://dx.doi.org/10.3390/electronics10111223]
129. Zhao, Z.; You, J.; Cui, Y. Freshness recognition of fruit and vegetable images using GANs series data augmentation. In: proceedings of the 2022 4th international conference on video, signal and image processing; IEEE: New York, NY, USA, 2022.
130. Adem, K; Közkurt, C. Defect detection of seals in multilayer aseptic packages using deep learning. Turk J Electr Eng Comput Sci; 2019; 27, pp. 4220-4230.
131. Medus, LD et al. Hyperspectral image classification using CNN: application to industrial food packaging. Food Control; 2021; 125, [DOI: https://dx.doi.org/10.1016/j.foodcont.2021.107962] 107962.
132. Momeny, M et al. Grading and fraud detection of saffron via learning-to-augment incorporated inception-V4 CNN. Food Control; 2023; 147,
133. Jahanbakhshi, A et al. A novel method based on machine vision system and deep learning to detect fraud in turmeric powder. Comput Biol Med; 2021; 136,
134. Ciocca, G; Napoletano, P; Schettini, R. CNN-based features for retrieval and classification of food images. Comput Vis Image Underst; 2018; 176, pp. 70-77.
135. Wang, Y et al. Deep learning in food safety and authenticity detection: an integrative review and future prospects. Trends Food Sci Technol; 2024; 104, pp. 396-406.
136. Maruthamuthu, MK et al. Raman spectra-based deep learning: a tool to identify microbial contamination. MicrobiologyOpen; 2020; 9, [DOI: https://dx.doi.org/10.1002/mbo3.1122] e1122.
137. Wu, T et al. Accurate prediction of salmon freshness under temperature fluctuations using the convolutional neural network long short-term memory model. J Food Eng; 2022; 334,
138. Tiwari, U. Neural network approach for risk assessment along the food supply Chain. Smart and sustainable food technologies; 2022; Singapore, Springer Nature: pp. 287-305.
139. Ziegler, P. Machine learning for inventory management: forecasting demand quantiles of perishable products with a neural network. MS Thesis, 2020.
140. Makridis, G; Mavrepis, P; Kyriazis, D. A deep learning approach using natural language processing and time-series forecasting towards enhanced food safety. Mach Learn; 2023; 112, pp. 1287-1313.
141. Khan, PW; Byun, Y-C; Park, N. IoT-blockchain enabled optimized provenance system for food industry 4.0 using advanced deep learning. Sensors; 2020; 20, 2990.
142. Du, Y et al. Foodborne disease risk prediction using multigraph structural long short-term memory networks: algorithm design and validation study. JMIR Med Inform; 2021; 9, [DOI: https://dx.doi.org/10.2196/29433] e29433.
143. Wang, J.; et al. Using convolutional neural networks to extract keywords and keyphrases: a case study for foodborne illnesses. In proceedings of the 2019 18th IEEE international conference on machine learning and applications (ICMLA); IEEE: New York, NY, USA, 2019.
144. Zhang, J; El-Gohary, NM. Semantic NLP-based information extraction from construction regulatory documents for automated compliance checking. J Comput Civ Eng; 2016; 30, 04015014.
145. Chen, J; Cao, H; Natarajan, P. Integrating natural language processing with image document analysis: what we learned from two real-world applications. Int J Doc Anal Recognit; 2015; 18, pp. 235-247. [DOI: https://dx.doi.org/10.1007/s10032-015-0247-x]
146. Blosseville, M.-J.; et al. Automatic document classification: natural language processing, statistical analysis, and expert system techniques used together. In proceedings of the 15th annual international ACM SIGIR conference on research and development in information retrieval; ACM: New York, NY, USA, 1992.
147. Khanbhai, M et al. Applying natural language processing and machine learning techniques to patient experience feedback: a systematic review. BMJ Health Care Inform; 2021; 28, pp. 1-11. [DOI: https://dx.doi.org/10.1136/bmjhci-2020-100262]
148. Bommarito, MJ; Katz, DM; Detterman, EM. LexNLP: natural language processing and information extraction for legal and regulatory texts. Research handbook on big data law; 2021; Cheltenham, Edward Elgar Publishing: pp. 216-227.
149. Agarwal, A; Jayant, A. Machine learning and natural language processing in supply chain management: a comprehensive review and future research directions. Int J Bus Insights Transform; 2019; 13, pp. 1-20.
150. Chauhan, C.; Sehgal, S. Sentiment analysis on product reviews. In proceedings of the 2017 international conference on computing, communication and automation (ICCCA); IEEE: New York, NY, USA, 2017.
151. Xiao, Y.; Qi, C.; Leng, H. sentiment analysis of amazon product reviews based on NLP. In proceedings of the 2021 4th international conference on advanced electronic materials, computers and software engineering (AEMCSE); IEEE: New York, NY, USA, 2021.
152. Thomas, K.K.; Anil, S.P.; Ebin Kuriakose, N.G (2019) Sentiment analysis in product reviews using natural language processing and machine learning. Int J Inf Syst Comput Sci
153. Park, SB; Jang, J; Ok, CM. Analyzing twitter to explore perceptions of Asian restaurants. J Hosp Tour Technol; 2016; 7, pp. 405-422.
154. Rose, SW et al. Perceptions of menthol cigarettes among twitter users: content and sentiment analysis. J Med Internet Res; 2017; 19,
155. Islam, N.; Akter, N.; Sattar, A. Sentiment Analysis on Food Review Using Machine Learning Approach. In Proceedings of the 2021 International Conference on Artificial Intelligence and Smart Systems (ICAIS); IEEE: New York, NY, USA, 2021.
156. Vatambeti, R et al. Twitter sentiment analysis on online food services based on elephant herd optimization with hybrid deep learning technique. Cluster Comput; 2024; 27, pp. 655-671. [DOI: https://dx.doi.org/10.1007/s10586-023-03970-7]
157. Zhang, D.; et al. IFoodCloud: A Platform for Real-Time Sentiment Analysis of Public Opinion about Food Safety in China. arXiv 2021, arXiv:2102.11033.
158. Song, C et al. An analysis of public opinions regarding take-away food safety: a 2015–2018 case study on Sina Weibo. Foods; 2020; 9, 511.
159. Du, C-J; Sun, D-W. Learning techniques used in computer vision for food quality evaluation: a review. J Food Eng; 2006; 72, pp. 39-55.
160. Zhang, B; Liu, Y; Wu, J; Yang, J; Yang, S. Principles, developments and applications of computer vision for external quality inspection of fruits and vegetables: a review. Food Res Int; 2014; 62, pp. 326-343.
161. Patel, KK; Patel, PR; Patel, PS; Patel, SJ. Machine vision system: a tool for quality inspection of food and agricultural products. J Food Sci Technol; 2012; 49, pp. 123-141.
162. Tretola, M.; Guida, F.; Vitiello, G.; D'Agostino, D.; Picerno, P.; Spagnuolo, M. (2017) former food products safety evaluation: computer vision as an innovative approach for the packaging remnants detection. J. Food Qual 1064580
163. Brosnan, T; Sun, D-W. improving quality inspection of food products by computer vision––a review. J Food Eng; 2004; 61, pp. 3-16.
164. Nithya, R; Santhi, B; Manikandan, R; Rahimi, M; Gandomi, AH. Computer vision system for mango fruit defect detection using deep convolutional neural network. Foods; 2022; 11, 3483. [DOI: https://dx.doi.org/10.3390/foods11213483]
165. Rong, D; Xie, L; Ying, Y. Computer vision detection of foreign objects in walnuts using deep learning. Comput Electron Agric; 2019; 162, pp. 1001-1010.
166. Son, G-J; Kim, D-H; Lee, J-H; Choi, W-S. U-net-based foreign object detection method using effective image acquisition system: a case of almond and green onion flake food process. Sustainability; 2021; 13, 13834. [DOI: https://dx.doi.org/10.3390/su132413834]
167. Haque, A.; Kotsis, S.V.; Caban, J.; Stone, J.M.; Liu, D.; Bakal, D. Towards Vision-Based Smart Hospitals: A System for Tracking and Monitoring Hand Hygiene Compliance. In Proceedings of the Machine Learning for Healthcare Conference; PMLR: 2017.
168. Gunasekaran, S. Computer vision technology for food quality assurance. Trends Food Sci Technol; 1996; 7, pp. 245-256.
169. Olsen, P; Borit, M. The components of a food traceability system. Trends Food Sci Technol; 2018; 77, pp. 143-149.
170. Abdullah, MZ; Shahid, S; Khan, S; Bhatti, S; Rahman, S. The applications of computer vision system and tomographic radar imaging for assessing physical properties of food. J Food Eng; 2004; 61, pp. 125-135.
171. Wang, J; Fu, P; Gao, RX. Machine vision intelligence for product defect inspection based on deep learning and Hough transform. J Manuf Syst; 2019; 51, pp. 52-60.
172. Patel, P; Doddamani, A. Role of sensor in the food processing industries. Int Arch Appl Sci Technol; 2019; 10, pp. 10-18.
173. Karim, AB; Alshehri, AA; Tariq, U; Salim, F; Javed, I. Monitoring food storage humidity and temperature data using IoT. MOJ Food Process Technol; 2018; 6, pp. 400-404.
174. Sidel, JL; Stone, H. The role of sensory evaluation in the food industry. Food Qual Prefer; 1993; 4, pp. 65-73.
175. Świąder, K; Marczewska, M. Trends of using sensory evaluation in new product development in the food industry in countries that belong to the EIT regional innovation scheme. Foods; 2021; 10, 446.
176. Vadivambal, R; Jayas, DS. Applications of thermal imaging in agriculture and food industry—a review. Food Bioproc Technol; 2011; 4, pp. 186-199.
177. Dai, Q; Liu, W; Zhou, X; Jiang, H; Zhang, J. Recent advances in data mining techniques and their applications in hyperspectral image processing for the food industry. Compr Rev Food Sci Food Saf; 2014; 13, pp. 891-905.
178. Kuzminov, I.; Dmitriev, K.; Koryakov, V. Mapping the Radical Innovations in Food Industry: A Text Mining Study. Higher School of Economics Research Paper No. WP BRP 80 2018.
179. Paramesha, K; Gururaj, HL; Jena, OP. Applications of machine learning in biomedical text processing and food industry. Machine learning for healthcare applications; 2021; Singapore, Springer: pp. 151-167.
180. Singh, SK; Jenamani, M. Cassandra-based data repository design for food supply chain traceability. VINE J Inf Knowl Manag Syst; 2021; 51, pp. 193-217.
181. Tian, F. A Supply chain traceability system for food safety based on HACCP, blockchain & internet of things. In: proceedings of the 2017 international conference on service systems and service management; IEEE: New York, NY, USA, 2017.
182. Sharma, R.; Parhi, S. A review on use of big data in warehousing to enhance accessibility of food. In: proceedings of the 2017 2nd international conference on communication and electronics systems (ICCES); IEEE: New York, NY, USA, 2017.
183. Krishnan, K. (2013) Data Warehousing in the age of big data. Newnes
184. de Assis Vilela, F; Silva, JA; Neves, PD. A non-intrusive and reactive architecture to support real-time ETL processes in data warehousing environments. Heliyon; 2023; 9, e15723.
185. Wardhani, FZD; Wiratama, J. Improving the quality of service: ETL implementation on data warehouse at pharmacy industry. Jurnal Tekno Kompak; 2024; 18, pp. 1-14. [DOI: https://dx.doi.org/10.33365/jtk.v18i1.3211]
186. Biancolillo, A; Marini, F; Pirovano, S. Data fusion strategies in food analysis. Data handling in science and technology; 2019; Amsterdam, Elsevier: pp. 271-310.
187. Borràs, E; Fernández, S; Fernández, S; Llorach, R; Rovira, J. Data fusion methodologies for food and beverage authentication and quality assessment–a review. Anal Chim Acta; 2015; 891, pp. 1-14.
188. Nychas, G-JE; Panagou, EZ; Mohareb, F. Novel approaches for food safety management and communication. Curr Opin Food Sci; 2016; 12, pp. 13-20.
189. Yu, Z; Xu, Y; Hu, Y; Wu, Z. Smart traceability for food safety. Crit Rev Food Sci Nutr; 2022; 62, pp. 905-916.
190. Chapman, J. Data accuracy and model reliability. In: proceedings of the BEPAC conference; Canterbury, UK, 1991.
191. Fan, W.; Geerts, F.; Jia, X. Improving data quality: consistency and accuracy. In: proceedings of the ACM SIGKDD international conference on knowledge discovery and data mining; ACM: New York, NY, USA, 2007.
192. Sheng, V.S.; Provost, F.; Ipeirotis, P.G. Get Another Label? Improving data quality and data mining using multiple, noisy labelers. In: proceedings of the 14th ACM SIGKDD international conference on knowledge discovery and data mining; ACM: New York, NY, USA, 2008.
193. Miller, R; Timmons, R; Hughes, L; Scott, A. How modeling and simulation have enhanced decision making in new drug development. J Pharm Sci; 2005; 32, pp. 185-197.
194. Hess, JD; Bacigalupo, AC. Enhancing decisions and decision-making processes through the application of emotional intelligence skills. Manag Decis; 2011; 49, pp. 710-721.
195. Verstichel, S; Van Landeghem, H; Van Hirtum, A; Verhoeff, J. Efficient data integration in the railway domain through an ontology-based methodology. Transp Res Part C: Emerg Technol; 2011; 19, pp. 617-643. [DOI: https://dx.doi.org/10.1016/j.trc.2010.10.003]
196. Patcha, A; Park, J-M. An overview of anomaly detection techniques: existing solutions and latest technological trends. Comput Netw; 2007; 51, pp. 3448-3470.
197. Rabin, N.; Averbuch, A. Detection of anomaly trends in dynamically evolving systems. In proceedings of the 2010 AAAI Fall symposium series; AAAI: Menlo Park, CA, USA, 2010.
198. Delbaere, M; Ferreira, R. Addressing the data aspects of compliance with industry models. IBM Syst J; 2007; 46, pp. 319-334.
199. Nguyen, Q; Franks, D; Harvey, H; Chan, M. Structured reporting of IR procedures: effect on report compliance, accuracy, and satisfaction. J Vasc Interv Radiol; 2018; 29, pp. 345-352.
200. Viswanathan, L.; Jindal, A.; Karanasos, K. Query and Resource optimization: bridging the gap. In: proceedings of the 2018 IEEE 34th international conference on data engineering (ICDE); IEEE: New York, NY, USA, 2018.
201. Ahmed, QW; Saeed, M; Shahid, M; Bibi, N; Younis, S. AI-based resource allocation techniques in wireless sensor internet of things networks in energy efficiency with data optimization. Electronics; 2022; 11, 2071.
202. Mazhar, SA; Ali, N; Khan, A; Qureshi, MA. Methods of data collection: a fundamental tool of research. J Integr Commun Health; 2021; 10, pp. 6-10.
203. Joshi, AP; Patel, BV. Data preprocessing: the techniques for preparing clean and quality data for data analytics process. Orient J Comput Sci Technol; 2021; 13, pp. 78-81.
204. Moore, R.C.; Lewis, W. Intelligent selection of language model training data. In proceedings of the ACL 2010 conference short papers; ACL: Uppsala, Sweden, 2010.
205. Evans, P. Scaling and assessment of data quality. Acta Crystallogr Sect D: Biol Crystallogr; 2006; 62, pp. 72-82.
206. Fredriksson, T.; Persson, L.; Ståhl, K.; Wähälä, K. Data Labeling: An empirical investigation into industrial challenges and mitigation strategies. In: proceedings of the international conference on product-focused software process improvement; Springer: Cham, Switzerland, 2020.
207. Tsioptsias, N.; Tako, A.; Robinson, S.P. Model validation and testing in simulation: a literature review. In: proceedings of the 5th student conference on operational research (SCOR 2016); Schloss Dagstuhl: Germany, 2016.
208. Bardenet, R.; Bousquet, O.; Brendel, M.; Kégl, B.; Lacoste-Julien, S. Collaborative Hyperparameter tuning. In: proceedings of the international conference on machine learning; PMLR: 2013.
209. Braiek, HB; Khomh, F. On testing machine learning programs. J Syst Softw; 2020; 164, [DOI: https://dx.doi.org/10.1016/j.jss.2020.110542] 110542.
210. Pal, A; Kant, K. Smart sensing, communication, and control in perishable food supply chain. ACM Trans Sens Netw (TOSN); 2020; 16, pp. 1-41.
211. Matindoust, S; Fattahi, A; Olyaee, S; Mahdavi, M; Dourandish, S; Ghasemi, I. Food quality and safety monitoring using gas sensor array in intelligent packaging. Sens Rev; 2016; 36, pp. 169-183. [DOI: https://dx.doi.org/10.1108/SR-07-2015-0115]
212. Chen, H; Liu, X; Wang, M; Wang, Q; Zhang, J. Nanomaterials as optical sensors for application in rapid detection of food contaminants, quality and authenticity. Sens Actuators B: Chem; 2021; 329, [DOI: https://dx.doi.org/10.1016/j.snb.2020.129135] 129135.
213. Adley, CC. Past, present and future of sensors in food production. Foods; 2014; 3, pp. 491-510.
214. Beć, KB; Grabska, J; Huck, CW. Miniaturized NIR spectroscopy in food analysis and quality control: promises, challenges, and perspectives. Foods; 2022; 11, 1465. [DOI: https://dx.doi.org/10.3390/foods11101465]
215. Wu, D; Sun, D-W. Advanced applications of hyperspectral imaging technology for food quality and safety analysis and assessment: a review—part I: fundamentals. Innov Food Sci Emerg Technol; 2013; 19, pp. 1-14.
216. El Matbouly, H; Nikbakhtnasrabadi, F; Dahiya, R. RFID near-field communication (NFC)-based sensing technology in food quality control. Biosensing and Micro-Nano devices: design aspects and implementation in food industries; 2022; Singapore, Springer: pp. 219-241. [DOI: https://dx.doi.org/10.1007/978-981-16-8333-6_9]
217. Pigini, D; Conti, M. NFC-based traceability in the food chain. Sustainability; 1910; 2017, 9.
218. Fernandez, CM; Lopez, A; Fernandez, J; Ruiz, S; Martinez, JL. Innovative processes in smart packaging. A systematic review. J Sci Food Agric; 2023; 103, pp. 986-1003.
219. McFarlane, I. Automatic control of food manufacturing processes; 2012; Berlin, Springer Science & Business Media:
220. Shahzad, A.; Zhang, K. An Integrated IoT-Blockchain Implementation for end-to-end supply chain. In: proceedings of the future technologies conference (FTC) 2020, Volume 2; Springer International Publishing: Cham, Switzerland, 2021.
221. Garg, D; Gupta, P; Kundu, A. Blockchain and supply chain management. The auditor’s guide to blockchain technology; 2022; Boca Raton, CRC Press: pp. 93-104.
222. Kharche, A.; Badholia, S.; Upadhyay, R.K (2024) Implementation of blockchain technology in integrated IoT networks for constructing scalable ITS systems in India. blockchain: Res Appl 100188
223. Rejeb, A; Rejeb, K; Keogh, JG. Potentials of blockchain technologies for supply chain collaboration: a conceptual framework. Int J Logist Manag; 2021; 32, pp. 973-994.
224. Singh, V; Sharma, SK. Application of blockchain technology in shaping the future of food industry based on transparency and consumer trust. J Food Sci Technol; 2023; 60, pp. 1237-1254.
225. Galvez, JF; Mejuto, JC; Simal-Gandara, J. Future challenges on the use of blockchain for food traceability analysis. TrAC Trends Anal Chem; 2018; 107, pp. 222-232. [DOI: https://dx.doi.org/10.1016/j.trac.2018.08.011]
226. Tayal, A; Choudhury, A; Ranjan, R. Blockchain-based efficient communication for food supply chain industry: transparency and traceability analysis for sustainable business. Int J Commun Syst; 2021; 34, [DOI: https://dx.doi.org/10.1002/dac.4696] e4696.
227. da Costa, TP; Silva, EM; Costa, MM. A systematic review of real-time monitoring technologies and its potential application to reduce food loss and waste: key elements of food supply chains and IoT technologies. Sustainability; 2022; 15, 614.
228. Romeo, L; Rahmani, AM; Younis, M; Alazab, M; Alazab, M. Internet of robotic things in smart domains: applications and challenges. Sensors; 2020; 20, 3355. [DOI: https://dx.doi.org/10.3390/s20123355]
229. Li, Z; Liu, J; Ma, J; Li, X. IoT-based tracking and tracing platform for prepackaged food supply chain. Ind Manag Data Syst; 2017; 117, pp. 1906-1916.
230. Abu, NS; Kharb, S; Bansal, A. Internet of things applications in precision agriculture: a review. J Robot Control (JRC); 2022; 3, pp. 338-347.
231. Nguyen, H.; Do, L. The adoption of blockchain in food retail supply chain: Case: IBM food trust blockchain and the food retail supply chain in Malta. Informatics 2018.
232. Misra, NN; Hossain, MA; Das, SK; Kumar, P. IoT, big data, and artificial intelligence in agriculture and food industry. IEEE Int Things J; 2020; 9, pp. 6305-6324.
233. Pinart, C.; Mora, E. Ethical AI/ML at Nestlé: From Vision to Strategy to Execution. J. AI, Robot. Workplace Autom. 2022, 1, 247–255.
234. Food Safety & Quality. Available online: https://www.cargill.com/foodservice/food-safety-quality (accessed on 1 August 2024).
235. Food Safety. Available online: https://www.zestlabs.com/challenges/food-safety/ (accessed on 1 August 2024).
236. Food Safety & Quality. Available online: https://www.tetrapak.com/sustainability/focus-areas/food-safety-quality (accessed on 1 August 2024).
237. Ecolab. Available online: https://connect.ecolab.com/s/foodsafety?language=en_US (accessed on 1 August 2024).
238. Automating Agriculture. Available online: https://www.riperobotics.com/#eve (accessed on 1 August 2024).
239. Clear Labs. Available online: https://www.clearlabs.com/ (accessed on 1 August 2024).
240. Lin, X; Zhao, Y; Li, C; Jiang, H; Zhao, W. Dynamic risk assessment of food safety based on an improved hidden Markov model integrating cuckoo search algorithm: a sterilized milk study. J Food Proc Eng; 2021; 44,
241. Rasheed, H. Consideration of cloud-web-concepts for standardization and interoperability: a comprehensive review for sustainable enterprise systems, AI, and IoT integration. J Info Technol Info; 2024; 3, pp. 2-20.
242. Sasikala, B; Sachan, S. Decoding decision-making: embracing explainable AI for trust and transparency. Exploring the frontiers of artificial intelligence and machine learning technologies; 2023; Singapore, Springer: pp. 31-40.
243. Zliobaite, I; Figueiredo, R; Giannopoulos, G. Next challenges for adaptive learning systems. ACM SIGKDD Explor Newsl; 2012; 14, pp. 48-55.
244. Sapienza, S. Big data, algorithms and food safety: a legal and ethical approach to data ownership and data governance; 2022; Cham, Springer Nature:
245. Adhikari, R.; Richards, D.; Scott, K. Security and privacy issues related to the use of mobile health apps. In: proceedings of the 2014 Australasian conference on information systems (ACIS); 2014.
246. Manning, L; Baines, R; Wang, H. Artificial intelligence and ethics within the food sector: developing a common language for technology adoption across the supply chain. Trends Food Sci Technol; 2022; 125, pp. 33-42. [DOI: https://dx.doi.org/10.1016/j.tifs.2022.04.025]
247. Ferrara, E. Fairness and bias in artificial intelligence: a brief survey of sources, impacts, and mitigation strategies. Sci; 2023; 6, 3.
248. Arellano, L.; Alcubilla, P.; Leguízamo, L. (2023) Ethical considerations in informed consent. In: Ethics-scientific research, ethical issues, artificial intelligence and education. IntechOpen, London
249. Zhuk, A. Artificial intelligence impact on the environment: hidden ecological costs and ethical-legal issues. J Digit Technol Law; 2023; 1, pp. 932-954.
250. Javaid, M; Haleem, A; Singh, RP; Kumar, S. substantial capabilities of robotics in enhancing industry 4.0 implementation. Cogn. Robot.; 2021; 1, pp. 58-75.
251. Handford, C.E.; McCann, M.; McGowan, C.; Green, S. Nanotechnology in the agri-food industry on the Island of Ireland: applications, opportunities and challenges. Institute for global food security at Queen’s university, Belfast, and the Teagasc Ashtown food research centre, Dublin, 2014.
252. Okoye, CC; Ezema, IJ; Nwosu, JC. Securing financial data storage: a review of cybersecurity challenges and solutions. Int J Sci Res Archive; 2024; 11, pp. 1968-1983.
253. Shah, V; Reddy Konda, S. Neural networks and explainable AI: bridging the gap between models and interpretability. Int J Comput Sci Technol; 2021; 5, pp. 163-176.
254. Boland, M; Alam, F; Bronlund, J. Modern technologies for personalized nutrition. Trends in personalized nutrition; 2019; Amsterdam, Elsevier: pp. 195-222.
255. Kumar, I; Sharma, S; Kumar, V; Meena, SK. Opportunities of artificial intelligence and machine learning in the food industry. J Food Qual; 2021; 2021, 4535567.
256. Díaz-Rodríguez, N; Fernández, JC; Martínez-Cámara, E. connecting the dots in trustworthy artificial intelligence: from AI principles, ethics, and key requirements to responsible AI systems and regulation. Inf Fusion; 2023; 99, [DOI: https://dx.doi.org/10.1016/j.inffus.2023.101896] 101896.
Copyright Springer Nature B.V. Jan 2025