Research And System Implementation Of Facial Expression Recognition Under Remote Network Monitoring

Facial expression is an important way for human to express emotion, by recognizing the facial expression, mentality can be analyzed in order to obtain valuable emotional information. To carry out the research of face recognition technology for remote network monitoring, we can use the Internet to get more extensive facial expression and dynamic change information in a timely manner. It has important theoretical significance and application value for the prevention of terrorist attacks against the extreme nationalism, intelligent traffic management, and so on.In this paper, based on the IP network camera to collect the facial expression image from the remote side, the fast optimization facial expression recognition algorithm is deeply studied. At the same time, using internet of things technology architecture for data transmission, setting up the remote monitoring system of MVC design pattern for facial expression recognition, integrating each function module,recognizing remote monitoring scene of facial expression.The main work completed in this thesis is listed as follows:(1)To study the basic principle and characteristics of the IP protocol, TCP protocol, HTTP protocol expansion based on IOT architecture. This system selects more reasonable IPv4/IPv6 transition mechanism, realizes the IPv4/IPv6 communication function module, understands the TCP/IP reference model, analyzes the HTTP message structure and request methods, and ensures the validity and consistency of data transmission.(2) A facial expression feature extraction algorithm based on 2D-LPCA is proposed to enhance the timeliness of the system. Local PCA algorithm is proposed for the recognition of facial expressions in the process of the existence of different expressions and similar expressions to be easily confused. LPCA algorithm improves the recognition rate of facial expressions, but increases the complexity of the algorithm when selecting the training samples, so combines with the 2D-PCA algorithm, proposes the complexity of the 2D-LPCA algorithm which increases the real-time performance of the algorithm. Experimental results show that the 2D-LPCA algorithm has the characteristics of simple principle, high recognition rate and fast speed.(3)The expression classifier based on Adaboost-Knn algorithm is designed and implemented to further improve the recognition rate and robustness of facial expressions. According to the nearest neighbor rule and the similarity measure criterion,in order to improve the recognition accuracy,The algorithm combines boosting classification to design Adaboost-Knn combination classifier. After experimental verification, by optimizing the parameters, the recognition classifier hasgood robustness and high rate.Design and construct a remote monitoring system for facial expression recognition. The system is based on the B/S architecture, which is implemented through the browser to access the system. Using MVC design pattern to isolate the data persistence layer, business layer and presentation layer, reduce the coupling of the system. Based on the SSH2 framework, with the help of JNI/JNA cross language to call opencv library system, by java applet technology to embed expression recognition algorithm, this system has four core modules: video surveillance,expression warehouse, data characteristics, system management.