A web-based pneumonia detection system using deep learning on chest X-ray image

Abstract

Pneumonia is a lung infection caused by bacteria, viruses, or fungi that can cause inflammation of the air cavities in one or both lungs. Pneumonia is diagnosed using chest X-rays, but results can take 1 to 3 hours from medical experts, causing delays in treatment for most patients who have pneumonia. Therefore, a web-based artificial intelligence pneumonia detection application was developed in this study. The website development consists of several stages: requirements analysis, system design, model selection and description, application development, and testing. The deep learning model uses a convolutional neural network algorithm with VGG16, which provides effective feature extraction for complex object recognition and has two output classes. The website framework was developed using Flask with MySql for the database. The simulation results show that the classifier model accuracy rate is 98%. The web application can classify chest X-ray images with 100% accuracy on a small test set consisting of five images: two normal and three showing pneumonia. Additionally, the inference model takes only 4.09 seconds to process the classification result. Thus, our proposed e-health web-based system for pneumonia detection is accurate and real-time.