Research And Application Of Deep Learning Of Online Learning Behavior Based On Face Recognition Technology

Compared with the traditional classroom teaching mode,online learning has many advantages such as being free from time and place restrictions.However,there are some outstanding problems in the learning supervision of learners in the current online learning system,such as how to verify the identity of learners,learners to learn how to detect the state,how to perceive the emotions of learners.This paper puts forward and realizes the method of depth perception and assessment for learners' learning behaviors from the aspects of face recognition,eye state detection and emotion perception.It provides a new perspective and a way for online learning behavior assessment.For online learning system,self-regulation and external supervision of learners provide technical support.The main work of the pape r is as follows:First of all,based on the analysis of the problems existing in the current online learning system in the supervision of learners,this paper proposes the establishment of an online learning behavioral depth perception model.The model includes three main functions:face recognition,eye state detection,and emotion perception.In view of the three problems of this model:face key location,eye state detection,emotional feature recognition,this paper puts forward the solution:Based on the cascade convolution neural network,the human face key points are accurately positioned by coarse positioning-fine positioning-fine positioning 3 steps.Eye detection and emotional feature recognition based on convolutional neural networks,characterization of eyes and mouth by determining eye opening and mouth radiation.A deep perception classification method based on CNN-SVM for online learning behavior is proposed,CNN is used to extract the features in the input image so they are not relevant.The linear kernel function support vector machine is used to classify the detected eye and mouth features,so that the classification effect is improved.Then,the data obtained by the three perception functions of face recognition,eye state detection,and emotion perception are used as evaluation data sources.Based on the C4.5 algorithm,this paper proposes an online learning behavior assessment method,builds an online learning behavior evaluation decision tree and conducts an online learning behavior assessment based on the established classification rules.Finally,based on the Face++ artificial intelligence open platform,this paper realizes the online learning behavior depth perception and evaluation system.The environment and tools for the development of the system are given,and the development and implementation of the function modules such as face recognition,eye state detection,emotion perception and learning behavior evaluation are introduced.The operation of the system is good.