A Research And Application Of Expression Recognition Based On RealSense In Online Education

Online education is an important way for education informatization and life-long education,and it is also one of the ways to develop individuality education.Relying on the Internet and computers and other technologies,online education have some important shifts and profound changes in educational concepts and educational models,but,there are still some deficiencies when the online education is practiced.Due to the separation of time and space of e-learning,the students and teachers can't achieve emotional communication during the learning process.These shortcomings will affect the learning efficiency of online education.So,the study of how to make up for the lack of emotions of learners in online education,has become an urgent issue in the field of online education.The emotional calculation of online education emphasizes the real-time acquisition and rapid feedback of learner's emotional information,which is diversity and complexity.The RealSense Technology is a mid-distance perception technology developed by Intel Corporation that can obtain facial feature data in real time.By analyzing the data of facial feature points,the teacher can obtain the expression descriptions of the students during the learning process in real time,and then interact with the students to make up for the lack of emotions caused by teacher-student separation,and meet the emotional interaction requirements of e-learning.The main work and innovations of this article include:Firstly,on the basis of researching and analyzing the existing expression databases,combined with the characteristics of online education,establishing an expression database based on real sense technology.Different from other databases,this database also includes the depth image of the corresponding frame,and the entire face of the three-dimensional feature point data.Secondly,through studying the theory of face detection based on optical flow method,and analyzing a large amount of expression data,an expression detection method based on local feature point analysis is proposed,which can efficiently detect facial expressions.The experimental results show that this method can significantly reduce the redundancy of expression data.Then,an expression recognition model based on real sense technology was designed and implemented.The model includes three steps:image preprocessing,feature extraction and classification.The features extract from the images include LBP feature and basic RGB color features.The extracted features are put into the Alexnet convolutional neural network model for deep recognition.The test results of the expression database show that this method can identify the facial expression better and meet the requirements of online education.Finally,a preliminary model of online emotional computing system is built.In order to verify the effectiveness of the system,a small-scale experimental test was conducted.The experimental results show that the system is effective.