Design And Implementation Of Emotion Recognition In Mobile Terminalbe System Based On Deep Learning

As a significant research topic in the fields of artificial intelligence,computer vision and human-computer interaction,facial expression recognition has been widely used in medical treatment,public security,education and other fields.Due to the rapid development of smart phones and the increasingly popular mobile applications represented by Android image recognition in the current era,the application of facial expression recognition technology in the field of Android image recognition is inevitable.However,most of the facial expression recognition in specific scenarios is currently done on the PC terminal and embedded development boards.Regardless of the high recognition rate,models are usually large in size,so it cannot meet the operation requirements of the mobile terminal.Therefore,the mobile-terminal facial expression recognition system was designed in this paper,which was applied to actual scenarios to get closer to the real life.This paper mainly includes researches on the design framework,recognition model and software implementation of the mobile-terminal facial expression recognition system.The main works was as follows:The framework of mobile-terminal facial expression recognition system was designed in this paper,which mainly includes three parts: video acquisition module,video transmission module and software module.The working principle of the system is as follows: it first obtains the surveillance video through the EZVIZ camera,then transmits the video through the router and the Internet cloud,and finally recognizes the remote facial expression through the mobile APP.Due to the limited operational capability of the mobile-terminal CPU,mini?Xception,a lightweight neural network architecture,was used for facial expression recognition in this paper.In addition,the training and parameter adjustment of mini?Xception network were carried out in this paper.Finally,a recognition model which can be transplanted to the mobile phone was established.It can be concluded from the experimental results that the mini?Xception network architecture can meet the needs of the mobile terminal.In this paper,Android Studio development tools were used to design and develop a mobile-terminal facial expression recognition software APP.The tool includes the following main functions: login,remote video acquisition,face detection,facial expression recognition,alarm prompt,etc.It calls the LBP cascade detector in the Open CV Android SDK toolkit to detect faces,and then calls the transplanted mini?Xception model to recognize facial expressions.The preliminary test results in the home environment and classroom indicate that the system designed in this paper has afairly good practicality.