A Research Of K-12 Student’s E-Learning Emotion Recognition Based On Video

In recent years,distance education has been sinking to K12 schools,and the audience of online education has been generalized in age.However,while online education provides cross-temporal support for learners and ensures resource sharing."emotional deficiency" is undermining K12 students’ interest in online learning.However,most of the current researches on emotion recognition in online learning are aimed at higher education or adult learners.Therefore,it is of great significance to carry out emotion recognition in online learning for K12 students.This paper focuses on the online learning emotions of K12 students.Based on the learning process video and the use of deep learning and other learning analysis technologies,this paper realizes the accurate automatic recognition of K12 students’ online learning emotions and improves the machine’s ability to perceive K12 students’ online learning emotions.The main contributions and features of this paper are as follows;Firstly,a affective database of K12 students is constructed.Since there is no video database related to emotion recognition of students at present,in order to carry out cognitive emotion research of K12 students,this paper collects physiological and image data of students under four different task situations through emotion induction experiment.Based on two sensors,camera and heart rate watch,the learning process of students is recorded,and two different modes of image and heart rate data are generated.Based on this,a bi-modal affective database is constructed.Secondly,a video-based emotion recognition method is designed.Focusing on the data source of video,this paper extract the emotional features of video by using the frame-attention network and distinguish the cognitive emotional state of K12 students.The experimental results show that the accuracy of this method for the recognition of pleasure,attention,confusion and boredom reaches 87.5%.In order to explore more emotional recognition methods for online learning,this paper tries to detect heart rate based on video,analyzes the correlation between video heart rate detection results and students’ emotional state,and puts forward a improvement scheme for students’emotional recognition.