The Research of Chinese Martial Arts Cross-Media Communication System Based on Deep Neural Network

Abstract

The spread of Chinese martial arts is crucial for the world to understand Chinese culture. If only relying on one transmission method, it will lead to the difference of transmission and its lack of certain real time. This will lead to differences in the understanding of Chinese martial arts, which is also not conducive to the spread of Chinese glorious culture. Cross-media communication technology can solve this communication difference problem very well. The deep neural network method was used to fuse relevant features of Chinese martial arts, and it also analyzes the feasibility of neural network technology in cross-media communication. At the same time, this study uses deep neural network to study the timeliness of Chinese martial arts in the process of cross-media communication. The research results show that the convolutional neural network can effectively extract the characteristics of Chinese martial arts and carry out effective dissemination. However, the hybrid convolutional neural network with temporal features has higher accuracy in extracting Chinese martial arts features. This hybrid convolutional neural network is more conducive to the dissemination of Chinese martial arts through cross-media technology, which can ensure its timeliness. The maximum error of deep neural network technology in predicting Chinese martial arts culture is only 2.67%. This part of the error comes from the action characteristics of Chinese martial arts culture, which shows that neural network technology has good feasibility.

This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

1. Introduction

In the history of the development of Chinese culture and civilization, Chinese martial art has a long history and it is a culture that has been passed down. It is not only a representation of sports, but also a representation of Chinese culture [1, 2]. For a country, the national image is important in the communication of the international community. Every country often has a culture as an image characteristic of national exchanges. The characteristics of national cultural image are often a long-term historical culture, which can represent the development experience and historical culture of this country. Chinese martial art is one of the characteristics of national image [3]. It not only represents the historical inheritance of this culture, but also reflects the development history of Chinese culture and the development experience of China [4]. For the country, national culture is a symbol that inspires national labor and struggle. It is also a belief of the people of the country. Chinese martial arts culture has experienced thousands of years of development history, and it can inspire the hard work and tenacious fighting spirit of the Chinese [5, 6]. As a Chinese citizen, we must respect our national image culture wherever we are. Chinese martial arts culture is also a way for the people of the world to understand China [7, 8]. A country or even a citizen wants to understand the development process of China, and Chinese martial arts culture is a good representative.

Since Chinese martial arts culture is a representative of Chinese cultural symbols, we need to pass on Chinese martial arts culture in various ways, so that people around the world can understand Chinese martial arts culture [9, 10]. In the early stage of development, China generally disseminated Chinese martial arts culture through newspapers and the Internet [11]. Whether it is a large-scale sports event or a foreign cultural exchange, China often regards Chinese martial arts as a cultural exchange product. For example, during the Beijing Summer Olympics in 2008, China showed the Chinese martial arts culture to the people of the world through a performance. Although this will not increase the number of gold medals in China, it is also an opportunity to show Chinese martial arts to the world. However, there are certain differences in the development of each country [12, 13]. Even some countries are relatively backward in development, which will lead to the lag of cultural transmission. Chinese martial arts culture is relatively special culture. In the process of dissemination, the dissemination of Chinese martial arts culture tends to produce certain differences [14]. This will lead to differences in the world’s understanding of Chinese culture. Chinese martial arts culture contains cultural features such as body movements, images, and language [15, 16]. It is difficult to spread the connotation of Chinese martial arts culture to the world just by relying on one communication method. Cross-media technology is a way of disseminating information through the Internet, newspapers, and television [17]. Chinese martial arts culture will improve the timeliness and accuracy of cultural communication through cross-media technology. However, it also has a certain error rate when collecting the characteristics of Chinese martial arts culture, which makes it difficult to ensure the effective dissemination of Chinese martial arts culture.

In order to solve the problem of collecting the characteristics of Chinese martial arts culture and the real-time dissemination of cross-media technology, deep neural network can be considered to be applied in the dissemination process of Chinese martial arts culture [18, 19]. Deep neural network technology is a technology that has been widely used in recent years. The basis of deep neural network is machine learning algorithm and simple neural network technology, which is mainly due to the limitation of computer performance and graphics processing unit (GPU) computing performance. The amount of parameter calculation of deep neural network is huge, which requires GPU technology for computing performance and high-performance computer storage technology. Deep neural network methods can extract deeper data features by increasing the number of network layers. With the rapid development of data volume in each field, deep neural network technology has been favored by people [20]. In the process of cross-media dissemination of Chinese martial arts culture, it not only needs to ensure the effective extraction of the characteristics of Chinese martial arts culture, but also requires the timeliness of dissemination. It is difficult to make sure that Chinese martial arts culture dissemination has better accuracy and timeliness by traditional network or television. The deep neural network method can effectively extract the characteristics and time characteristics of Chinese martial arts culture, which can ensure the accuracy and timeliness of Chinese martial arts culture in the process of cross-media dissemination.

This study uses the deep neural network method to study the accuracy and timeliness of Chinese martial arts culture in the process of cross-media communication. This paper mainly introduces five aspects. The first part mainly explains the development of Chinese martial arts culture and the necessity of its dissemination. It also introduces the defects in the dissemination of Chinese martial arts culture. The second part mainly introduces the research status of Chinese martial arts culture. The third part mainly introduces the system composition and deep neural network method of the cross-media communication process of Chinese martial arts culture. The deep neural network technology used in this study is convolutional neural network (CNN) and hybrid neural network CNN-LSTM technology. The fourth part uses different statistical parameters to introduce the accuracy and timeliness of deep neural network technology in the process of Chinese martial arts culture cross-media dissemination. The fifth part is the summary part of the article.

2. Related Work

Chinese martial arts culture is more important to Chinese cultural image, and it will have different characteristics in the process of dissemination. Many researchers have also conducted related research on Chinese martial arts culture or the characteristics of martial arts culture. Peng and Zhang [21] believed that, with the increasing frequency of international exchanges, cultural exchanges between countries will gradually increase. It uses the comparative analysis method and the variation method to study the characteristics of Chinese martial arts culture. It also analyzes the influence of Chinese martial arts culture on the communication of Chinese martial arts art. Based on this method, he established a reliable model to evaluate the influence of Chinese martial arts culture on the spread. The results of the study indicate that propaganda will affect the communication of Chinese martial arts culture and it will strengthen the international cultural exchange. Ying [22] regarded traditional Chinese martial arts culture as an excellent traditional culture, but with the development of economy it has faded out of people vision. He used numerical analysis and data processing methods to establish a model of influencing factors of Chinese martial arts. He also studied the satisfaction of Chinese martial arts culture and art masses by using goal planning method and correlation analysis method. Through research, he suggested that Chinese martial arts should receive greater publicity, so that the masses can truly understand the connotation of Chinese martial arts. Guo and Fu [23] analyzed the development background of the exposure of martial arts culture in Guangdong. They also analyzed the manifestations of various martial arts cultures and the dissemination of traditional martial arts art and cultures. The results of the study show that the ideology of Guangdong martial arts culture and art originates from traditional Chinese philosophy. The values and symbols of martial arts culture are the important support for the development of Guangdong martial arts culture. Richards [24] researched the art form of Chinese martial arts and the contradiction between the ideological concept of Chinese martial arts art and Chinese martial arts art with film as the theme. The findings suggest that martial arts movie stars can influence the ideological change between contemporary and traditional culture and they can deepen the understanding of Chinese martial arts culture. Xu [25] believed that the traditional culture of Chinese martial arts represents China’s long history, and it can be regarded as a symbol of Chinese culture by foreigners. He also believes that it has an important relationship with Chinese martial arts culture and the spiritual connotation of spreading martial arts culture. This research takes Chinese hosting of the Chinese martial arts Art and Culture Festival as the research object. He uses the method of time series to establish an international art development model for the development of Chinese martial arts culture. The research results show that Chinese martial arts culture has been widely spread around the world, and Chinese martial arts culture and art are at a low level in the world. An and Hong [26] proposed that Maurice Halbwachs collective memory theory is helpful to understand Chinese understanding of martial arts art culture. They used interdisciplinary knowledge to build a memory spectrum of Chinese martial arts culture. Lv [27] mainly conducted related martial arts culture research based on the competition of Chinese martial arts culture. He found that the strategies and actions of Chinese martial arts culture have a corresponding relationship with each other. He used the competitive decision-making algorithm to build the decision-making algorithm of Chinese martial arts. He also studied the decision-making algorithm of Chinese martial arts culture using neural network technology and novel gradient descent method with added momentum and BP neural network method. This study compares the choice of step size as well as the choice of weight size and learning rate. The research results show that the model proposed in this study improves the scale and running time of Chinese martial arts culture decision-making, and it can help professionals to formulate scientific decision-making strategies. From the above literature review, it can be found that researchers have done a lot of research on the development and dissemination of Chinese martial arts culture; it involves traditional mathematical methods and simple neural network methods. It mainly refers to shallow neural network technology, such as BP neural network technology. These neural network techniques can only map the relationship between input and output. These studies do not involve deep neural network methods for Chinese martial arts cultural propagation tasks. The CNN and CNN-LSTM hybrid neural network technology used in this study can extract the characteristics and temporal characteristics of Chinese martial arts culture, which can help relevant personnel to discover the potential characteristics. It is helpful for the dissemination of Chinese martial arts culture.

3. An Application of Deep Neural Network in Chinese Martial Arts

3.1. The Introduction to Deep Neural Network

A deep neural network is a neural network technique that goes deeper into layers. It can handle the relationship between more complex data, because the amount of data has grown rapidly with the rapid development of economic globalization. The increase in the amount of data can no longer be handled by a simple neural network. BP neural network is the basis of neural network technology; it has the most basic perceptron and backpropagation mechanism. Deep neural network technology also has the basic process of backpropagation mechanism and perceptron. However, deep neural network technology allows deeper networks to perform related prediction operations through weight sharing. The cross-media communication technology of Chinese martial arts culture involves many features, which requires a deeper network level for feature extraction and prediction tasks. Therefore, deep neural network technology is a necessary method to more accurately predict and extract the characteristics and timeliness in the process of Chinese martial arts culture dissemination [28, 29]. Simple neural network technology does not have a weight sharing mechanism, which will increase the amount of parameter computation for the extraction of Chinese martial arts cultural features. There will be many features involved in the process of Chinese martial arts cultural cross-media dissemination. It is difficult for simple neural network technology to effectively extract this potential information.

3.2. The System Design and CNN Algorithm

Cross-media communication can effectively promote the accuracy and timeliness of the dissemination of Chinese martial arts culture, which can ensure the status of Chinese martial arts culture in world culture and the dissemination of Chinese historical information. This research mainly uses the deep neural network method to study the spread of Chinese martial arts culture. Figure 1 shows the system design scheme of Chinese martial arts culture cross-media dissemination. First of all, this research needs to collect the action features, image features, and language features of Chinese martial arts culture, which is to prepare for the deep neural network technology. These data need to go through the data processing stage, and then the processed Chinese martial arts cultural characteristic data is input into the convolutional neural network (CNN) and the long short-term memory neural network (LSTM) [30, 31]. After the relevant features of Chinese martial arts culture are extracted by CNN and LSTM, the data is output after passing through the output layer. These output data are feature data processed by deep neural network technology. These data are the Chinese martial arts culture data that need to be disseminated. Cross-media communication can be carried out in the form of television, newspapers, and the Internet, and they are interdependent [32]. In this study, the characteristics processed by deep neural network technology will be used to spread Chinese martial arts culture through television, newspapers, and the Internet. These characteristics can be propagated in these three media technologies simultaneously, and there is also a mutual connection between them.

[figure(s) omitted; refer to PDF]

The deep neural network technology used in this study includes CNN, which is used to extract the spatial characteristics of Chinese martial arts culture, and this technology can also reduce the amount of parameter computation of the deep neural network. Figure 2 shows the computational process of CNN in the extraction of Chinese martial arts cultural features. In Figure 2, the large cube represents the process of feature extraction by the convolutional layer, and each face of the cube represents a feature. The small cube represents the operation process of the pooling layer. These features first need to go through the convolutional layer of CNN to extract the main features; this process will reduce the data volume of features with lower weights. These features are then sampled by pooling layers and nonlinear data processing techniques with activation functions. Finally, the CNN will output the predicted data value through the output layer. The output data will be compared to the actual tag data value. This will involve the backpropagation mechanism of CNN and the error calculation process.

[figure(s) omitted; refer to PDF]

Whether it is a simple BP neural network or a complex deep neural network, it will involve two processes of forward propagation and backpropagation. These two processes involve complex derivation operations. But the derivation operation is solved by automatic differentiation in deep neural networks; equations (1) and (2) show the process of the derivation operation.$$\begin{array}{}\text{(1)}& \Delta {\omega }_{ji}=-\eta \frac{\partial E}{\partial {\omega }_{ji}},\text{(2)}& \Delta u_{ij}=-\eta \frac{\partial E\partial }{\partial u_{ij}}.\end{array}$$

In the process of backpropagation, CNN will involve many error calculations, which also involve the error calculation of weights and biases. Equations (3) and (4) show a representation of the error calculation.$$\begin{array}{}\text{(3)}& E=\frac{1}{2}{\displaystyle \sum _{k=1}^m{d_k-fnet{w}_k}^2}=\frac{1}{2}{\displaystyle \sum _{k=1}^m{d_k-f{\displaystyle \sum _{j=0}^n{\omega }_{jk}{y}_j}}^2},\text{(4)}& E=\frac{1}{2}{\displaystyle \sum _{k=1}^m{d_k-fnet{w}_k}^2}=\frac{1}{2}{\displaystyle \sum _{k=1}^m{d_k-f{\displaystyle \sum _{j=0}^n{\omega }_{jk}{y}_j}}^2}=\frac{1}{2}{\displaystyle \sum _{k=1}^m{d_k-f{\displaystyle \sum _{j=0}^n{\omega }_{j\kappa }f{\displaystyle \sum _{j=0}^q{u}_{ij}{\chi }_i}}}^2}.\end{array}$$

Error calculations can take many forms, such as mean square error, square root error, etc. Mean squared error is a commonly used method for calculating predicted and actual values. Equation (5) shows the principle of the mean square error calculation.$$\begin{array}{}\text{(5)}& L=\text{MSE}{q}^{\text{real}},q^{\text{pre}}=\frac{1}{nm}{\displaystyle \sum _{k=1}^N{\displaystyle \sum _{j=1}^M{q_{kj}^{\text{real}}-q_{kj}^{\text{pre}}}^2}}.\end{array}$$

The activation function is the key function for nonlinear processing, and equation (6) shows the calculation criterion of the activation function.$$\begin{array}{}\text{(6)}& Sx=\frac{1}{1+e^{-x}}.\end{array}$$

3.3. The Introduction to LSTM Algorithm and Data Processing

The CNN algorithm can only deal with spatial features very effectively, but it is not good at dealing with temporal features. LSTM can memorize some historical information, it has been widely used in the field of speech recognition, and it can handle temporal features very well. The dissemination of the characteristics of Chinese Wushu culture also requires the extraction of temporal characteristics, because there are certain differences in the temporal characteristics of Chinese Wushu culture in different periods. Therefore, this study adopts the method of CNN-LSTM hybrid deep neural network to explore the feasibility and accuracy of cross-media communication of Chinese martial arts culture. Figure 3 shows the application process of LSTM in the cross-media communication of Chinese martial arts culture. The input data of LSTM comes from the output data of CNN. This study does not directly provide LSTM with time series data of Chinese martial arts culture. LSTM and CNN have great differences in structure, and its biggest feature is that it has some gate structures. These gate structures can not only input the data, but also input historical state information. It can ensure the preservation of the temporal characteristics of Chinese martial arts culture. The data required by CNN is the data type of input and output, and the data required by LSTM is in the form of time series. The output data of CNN will be input to the input layer of LSTM in the form of time series.

[figure(s) omitted; refer to PDF]

The first layer of LSTM is the input gate structure, which can not only accept the input data of the current state, but also input historical information data. It can give different weights to the current state data and historical information data, which can ensure the integrity of the data and it can retain the time characteristics. Equation (7) shows the calculation criteria for the input gate.$$\begin{array}{}\text{(7)}& f_t=\sigma w_f•h_{t-1},P_t+b_f.\end{array}$$

The function of the forget gate is to filter the historical state information, in which a part of the features with a relatively large weight is retained, and a part of the features with a relatively small weight is filtered. This method not only ensures the preservation of the main temporal features, but also ensures that the amount of parameters is not too large. Equation (8) shows how the forget gate is calculated. Equation (9) shows the calculation process of the activation function.$$\begin{array}{}\text{(8)}& i_t=\sigma {\omega }_i•h_{i-1},P_t+b_i,\text{(9)}& {\tilde{C}}_t=\tanh w_c•h_{t-1},P_t+b_c.\end{array}$$

Equation (10) shows the calculation method of the refresh gate, which assigns different weights to the historical state information and the current state information.$$\begin{array}{}\text{(10)}& {\stackrel{⟶}{C}}_t=f_t\times {\stackrel{⟶}{C}}_{t-1}+i_t\times {\stackrel{⟶}{C}}_t.\end{array}$$

An output gate is a way to output a time series. Equation (11) shows the calculation of the output gate, which preserves some of the characteristics of the historical state information. Equation (12) shows how the data of the output gate goes through the nonlinear operation of the activation function.$$\begin{array}{}\text{(11)}& O_t=\sigma w_o•{\stackrel{⟶}{h}}_{t-1},P_t+b_o.\text{(12)}& {\stackrel{⟶}{h}}_t=O_t\times \tanh {\stackrel{⟶}{C}}_t.\end{array}$$

Data processing is a critical step before deep learning training and testing. The characteristics of the cross-media dissemination process of Chinese martial arts culture will involve the movement, image, and language characteristics of Chinese martial arts culture, and these characteristics will be quite different in the collection process. If this study directly uses these data as the input data of CNN, this may cause divergence and lower accuracy in the training process of CNN and LSTM. Therefore, before the training of CNN and LSTM, we need to normalize the characteristic data such as movements and images of Chinese martial arts culture.

4. Result Analysis and Discussion

The data collected in this study come from the relevant data of Chinese martial arts culture performances during the Beijing Olympic Games, which includes the movement characteristics, image characteristics, and language characteristics of Chinese martial arts culture. The cross-media communication process of Chinese martial arts culture will involve the movement, image, and language characteristics of martial arts. The accuracy and timeliness of these characteristics will affect the spread of Chinese martial arts culture. Therefore, this study conducts a related prediction study on three characteristics of Chinese martial arts culture through deep neural network technology. In this study, two methods, CNN and CNN-LSTM, were used to conduct related research. Figure 4 shows the prediction errors of three characteristics of Chinese martial arts culture using the CNN method. In general, the prediction errors of the characteristics of the three Chinese martial arts cultures are all within 3%, which can well guarantee the accuracy of Chinese martial arts culture in the process of cross-media dissemination. The largest error is only 2.78%, and this part of the error comes from the movement characteristics of Chinese martial arts culture. Action characteristics are difficult to be effectively captured by cross-media technology, and it is also difficult to ensure the accuracy of Chinese martial arts cultural characteristics. However, CNN effectively captures the action characteristics of Chinese martial arts culture. The smallest error is only 1.94%, and this part of the error comes from the image features of Chinese martial arts culture.

[figure(s) omitted; refer to PDF]

The parameters of CNN mainly include 5 layers of convolution layers and pooling layers, the number of filters is 128, and the stride is set to 1. The number of layers of CNN-LSTM includes 5 layers of CNN layers and 5 layers of LSTM layers, and the last layer is fully connected layer placed after the LSTM layer. Chinese martial arts culture is a characteristic with time characteristics. In order to verify the superiority of CNN-LSTM, this study compares the error of CNN method and CNN-LSTM in predicting the cultural characteristics of Chinese martial arts. Figure 5 shows the prediction errors of Chinese martial arts cultural features by the CNN-LSTM method. From Figure 5, it can be intuitively seen that the prediction errors of the three Chinese martial arts cultural characteristics are relatively small compared to the errors of the CNN method, which shows that the CNN-LSTM method is more superior in predicting the Chinese martial arts cultural characteristics. The error distribution of the three features such as action and image is also consistent with the error obtained by the CNN method. The largest error is only 2.67%. This part of the error is also derived from the movement characteristics of Chinese martial arts culture, and this part of the error has been reduced. Language features are data with obvious temporal characteristics, and this part of the error has been reduced to 1.72%.

[figure(s) omitted; refer to PDF]

The largest error in the cultural characteristics of Chinese martial arts comes from the prediction of action characteristics. This study selects the movement characteristics of Wushu alone for analysis. Figure 6 shows the prediction error distribution map of the action features of Chinese martial arts culture using the CNN-LSTM method. It can be seen from Figure 6 that all action errors are distributed within 4%. The prediction error distribution of the action characteristics of Chinese martial arts culture is relatively uniform. There is a large error difference on the left side of the image, but a small error distribution on the right side of the image. This may be due to the superior performance of CNN-LSTM gradually revealed over time. The action features of Chinese martial arts culture are features that have a clear relationship with temporal features, which requires good performance of CNN-LSTM. In general, the action features of Chinese martial arts culture will be better captured by the CNN-LSTM method. If the action features can be captured accurately and in real time, it will be beneficial to the collection and dissemination of cross-media technology.

[figure(s) omitted; refer to PDF]

Image feature is a feature with relatively small prediction error in the process of cross-media communication of Chinese martial arts culture. Figure 7 shows the distribution of prediction errors for image features of Chinese martial arts culture. The prediction error of image features is different from the distribution of action prediction error, and its distribution is relatively uniform. Moreover, the image prediction errors are basically within 2%. This is more favorable for cross-media communication. The images of Chinese martial arts culture can be spread through newspapers or TV, and this prediction error can well preserve the basic characteristics of Chinese martial arts culture. The distribution of the larger gradient of the image prediction error is mainly in the middle area of the image, which is because the middle area of the image has more characteristics of Chinese martial arts culture. However, in general, image feature prediction of Chinese martial arts culture can ensure effective dissemination through newspapers or television.

[figure(s) omitted; refer to PDF]

The language feature of Chinese martial arts culture is one of the most temporal features. Whether it is spread through TV or the Internet, it needs to ensure the accuracy of the language characteristics of Chinese martial arts culture. Figure 8 shows the distribution curve of language feature prediction of Chinese martial arts culture. Blue represents the actual language feature data, and red represents the predicted language feature data. In general, the predicted values of language features are in good agreement with the actual data values. In the early stage of language prediction, there is a large error; that is, the area of the green area is relatively large. But over time, the prediction error of the language features of Chinese martial arts culture becomes smaller. Figure 9 shows the scatter plot of the prediction errors of Chinese martial arts culture and language features. The prediction error of all the language features of Chinese martial arts culture is within 2%, which further illustrates the reliability of CNN-LSTM for the cross-media communication of Chinese martial arts culture. Larger language prediction errors still appear in the early stage of Chinese martial arts language features.

[figure(s) omitted; refer to PDF]

5. Conclusions

Chinese martial arts culture not only represents a kind of sports strength, but also reflects China’s history and culture as well as China’s development history. For the country, Chinese martial arts culture is a national cultural symbol. For citizens, Chinese martial arts culture is an image force that inspires their struggle. The accurate and real-time dissemination of Chinese martial arts culture is an important aspect for both the country and the citizens. There are big differences in the development of science and technology in each country, which will result in differences in the dissemination of Chinese martial arts culture. Cross-media technology is a communication technology that integrates multiple media, which can efficiently spread the movements, images, and language of Chinese martial arts. However, the feature extraction of Chinese martial arts culture is difficult for cross-media technology, and it is difficult to ensure the accuracy and real-time nature of Chinese martial arts cultural features. This study uses deep neural network technology to conduct related research and discussion on this issue.

This study uses two methods, CNN and CNN-LSTM, to conduct prediction research on the relevant characteristics of Chinese martial arts culture in the process of cross-media communication. The research results show that both CNN and CNN-LSTM methods can predict the movement, image, and language features of Chinese martial arts culture well. However, CNN-LSTM has better performance compared to CNN method because it can capture the temporal features of Chinese martial arts culture. The largest prediction error is only 2.67%, and this part of the error comes from the movement characteristics of Chinese martial arts culture. The smallest error is 1.72%, which is a relatively reliable prediction result for the cross-media communication of Chinese martial arts culture. As for the movement prediction error of Chinese martial arts culture, its main error is mainly distributed in the early stage of martial arts culture dissemination. Over time, this part of the error gradually decreases. This further shows that there is a strong time characteristic in the process of cross-media dissemination of Chinese martial arts culture. The distribution of language feature predictions of Chinese martial arts culture is relatively uniform. In general, deep neural network technology has strong accuracy and feasibility in the process of cross-media dissemination of Chinese martial arts culture.

Acknowledgments

The authors acknowledge Tianjin Philosophy and Social Science Planning Project, Research on the Theory and Practice of Tai Chi Intercultural Communication (no. TJTYQN20-001).

References

[1] V. Romanenko, L. Podrigalo, W. J. Cynarski, R. Olga, K. Lesia, V. Goloha, I. Robak, "A comparative analysis of the short-term memory of martial arts athletes of different levels of sportsmanship," Journal of martial arts anthropology, vol. 20 no. 3, pp. 18-24, 2020.

[2] A. H. Gross, R. Feldman, O Zagoory-Sharon, Y. Rassovsky, "Hormonal reactivity during martial arts practice among high-risk youths," Psychoneuroendocrinology, vol. 121,DOI: 10.1016/j.psyneuen.2020.104806, 2020.

[3] R. B. Santos-Junior, A. C. Utter, S. R. McAnulty, R. B. R. Bittencourt Bernardi, C. F. Buzzachera, E. Franchini, T. P. Souza-Junior, "Weight loss behaviors in Brazilian mixed martial arts athletes," Sport Sciences for Health, vol. 16 no. 1, pp. 117-122, DOI: 10.1007/s11332-019-00581-x, 2020.

[4] Q. S. Han, M. Theeboom, D. Zhu, "Chinese martial arts and the olympics: analysing the policy of the international wushu federation," International Review for the Sociology of Sport, vol. 11 no. 1, pp. 10-21, DOI: 10.1177/1012690220957177, 2020.

[5] C. Hongtu, Z. Cai, "The formal evolution of Chinese martial arts action movies and the spread of core values," The frontiers of society, science, and technology, vol. 2 no. 4, 2020.

[6] T. Aung, S. Goetz, J. Adams, C. McKenna, C. Hess, S. Roytman, J. T. Cheng, S. Zilioli, D. Puts, "Low fundamental and formant frequencies predict fighting ability among male mixed martial arts fighters," Scientific Reports, vol. 11 no. 1,DOI: 10.1038/s41598-020-79408-6, 2021.

[7] S. J. Stellpflug, W. H. Menton, R. C. LeFevere, "Analysis of the fight-ending chokes in the history of the Ultimate Fighting Championship mixed martial arts promotion," The Physician and Sportsmedicine, vol. 31 no. 1,DOI: 10.1080/00913847.2020.1866958, 2020.

[8] Y. Tian, "Research on the training path of innovative talents of martial arts and traditional national sports in the new era," International journal of education and economics, vol. 3 no. 4, 2020.

[9] H. Miyata, D. Kobayashi, S. Akifumi, H. Motoike, S. Akatsuka, "Mindfulness and psychological health in practitioners of Japanese martial arts: a cross-sectional study," BMC sports science, medicine & rehabilitation, vol. 12 no. 1,DOI: 10.1186/s13102-020-00225-5, 2020.

[10] M. Theeboom, D. Zhu, J. Vertonghen, "“Wushu belongs to the world,” But the gold goes to China…: the international development of the Chinese martial arts," International Review for the Sociology of Sport, vol. 52 no. 1,DOI: 10.1177/1012690215581605, 2017.

[11] P. Pawelec, P. Świder, W. J. Cynarski, "Martial arts tourism: meta-analysis of publications from 2005-2020," Sustainability, vol. 12 no. 17,DOI: 10.3390/su12177069, 2020.

[12] A. A. Figueiredo, W. Błach, Z. Bujak, R. J. Maroteaux, W. J. Cynarski, "Martial arts tourism of the “Europe-Far east” direction, in the opinion of grand masters," Sustainability, vol. 12 no. 21,DOI: 10.3390/su12218857, 2020.

[13] Y. Jiang, D. Hu, J. Hu, "Based on goal programming method of traditional Chinese martial arts culture protection and inheritance," Journal of Computational and Theoretical Nanoscience, vol. 13 no. 12,DOI: 10.1166/jctn.2016.6203, 2016.

[14] J. Zhang, "A cultural discourse analysis to Chinese martial arts movie in the context of glocalization: taking crouching tiger, hidden dragon and hero as cases," Advances in Language and Literary Studies, vol. 10 no. 3, pp. 32-41, DOI: 10.7575/aiac.alls.v.10n.3p.32, 2019.

[15] Q. Yajun, "Evolution and prospect of Chinese martial arts during the rising course of China," Proceedings of the 2016 2nd International Conference on Economics, management engineering and education technology, pp. 1121-1123, .

[16] S. S. M. Fong, J. S. M. Chan, Y. H. Bae, T. T. T Yam, L. M. Y. Chung, A. W. W. Ma, R. Kuisma, "Musculoskeletal profile of middle-aged Ving Tsun Chinese martial art practitioners: a cross-sectional study," Medicine, vol. 96,DOI: 10.1097/MD.0000000000005961, 2017.

[17] I. Ma, "Between historicity and fictionality: Xiang kairan, martial arts fiction, and Chinese narrative tradition," Modern Chinese Literature and Culture, vol. 31 no. 1, pp. 229-263, 2019.

[18] Z. Ai, "Quantitative CT study of martial arts sports injuries based on image quality," Journal of Visual Communication and Image Representation, vol. 60 no. 4, pp. 417-425, DOI: 10.1016/j.jvcir.2019.03.013, 2019.

[19] J. Yang, "Martial arts cinema in civil society of postcolonial Hong Kong: kung fu hustle and the grandmaster," Critical Arts, vol. 34 no. 6, pp. 28-40, DOI: 10.1080/02560046.2020.1826551, 2020.

[20] Y. Jia, M. Theeboom, D. Zhu, "Teaching traditional Chinese martial arts to con-temporary Chinese youth—a qualitative study with youth wushu coaches in China," Archives of budo, vol. 16 no. 7, 2020.

[21] X. L. Peng, Q. W. Zhang, "Influence factors and improvement suggestions of Chinese wushu communication in Chinese martial arts," Ekoloji, vol. 28 no. 107, pp. 2755-2760, 2019.

[22] L. Ying, "New ecology: Chinese traditional martial arts culture protection and carry forward game analysis," Ekoloji, vol. 28 no. 107, pp. 2773-2778, 2019.

[23] Y. Gao, C. K. Fu, "The culture of Guangdong martial arts," Sage Open, vol. 9 no. 3, 2019.

[24] S. Richards, "The patriotic narrative of Donnie Yen: how martial arts film stars reconcile Chinese tradition and modernity," Celebrity Studies, vol. 10 no. 2, pp. 276-284, DOI: 10.1080/19392397.2019.1601427, 2019.

[25] F. Xu, "Dissemination and inheritance of wushu spirit in the context of cultural globalization and it social environment," Ekoloji, vol. 28 no. 107, pp. 2867-2772, 2019.

[26] L. An, F. Hong, "Body center dot experience center dot imagination: the collective memory of Chinese martial arts," International Journal of the History of Sport, vol. 35 no. 15, pp. 1588-1602, DOI: 10.1080/09523367.2019.1620733, 2018.

[27] H. Lv, "Martial arts competitive decision-making algorithm based on improved BP neural network," Journal of healthcare engineering, vol. 7 no. 10, 2021.

[28] H. Fu, J. Cheng, Y. Xu, C. Zhang, D. W. K. Wong, J. Liu, X. Cao, "Disc-aware ensemble network for glaucoma screening from fundus image," IEEE Transactions on Medical Imaging, vol. 37 no. 2, pp. 2493-2501, DOI: 10.1109/tmi.2018.2837012, 2018.

[29] K. A. Mills, "What are the threats and potentials of big data for qualitative research?," Qualitative Research, vol. 18 no. 6, pp. 591-603, DOI: 10.1177/1468794117743465, 2018.

[30] A. Mohamed, M. K. Najafabadi, Y. B. Wah, E. A. K. Zaman, R. Maskat, "The state of the art and taxonomy of big data analytics: view from new big data framework," Artificial Intelligence Review, vol. 53 no. 2, pp. 989-1037, DOI: 10.1007/s10462-019-09685-9, 2020.

[31] Y. Wang, F. Yang, "A fractional-order CNN hyperchaotic system for image encryption algorithm," Physica Scripta, vol. 96 no. 3,DOI: 10.1088/1402-4896/abd50f, 2021.

[32] A. Zhang, N. Lv, "Research on the impact of big data capabilities on government’s smart service performance: empirical evidence from China," IEEE Access, vol. 9 no. 4,DOI: 10.1109/access.2021.3056486, 2021.