Design And Implementation Of Emotion Recognition System Based On Virtual Reality And EEG

With the rapid development of Virtual Reality(VR)technology,VR-based applications and research are also emerging,this paper combined with EEG and VR technology applied to emotional research.It is important to identify the emotional state of VR-users by EEG.On the one hand,the strong sense of immersion of VR can induce the wearer to produce the expected emotional state and recognize the emotional state through the EEG signal,which has a wide range of psychotherapy Application space;on the other hand,to identify the VR-device wearer’s emotional state in the VR-shopping,VR-games,VR-education and other areas of personalized service has a broad application prospects.This paper studies how to identify the emotional state of the virtual reality device by the EEG signal,and designs the emotion recognition system based on VR and EEG.The main work of this paper is as follows:1)This paper presents the use of VR-videos as a stimulus material to induce subjects to produce the expected emotional state of the method.There are less effect and other issues that when we use the videos,music,pictures as the stimulus material,in this article,we choose 8 VR-videos stimulation material to produce calm-excited(or nonarousal-arousal)emotions which rated by 30 subjects.A total of 240 experimental samples were obtained to establish a database of EEG signals and emotional states with the SAM emotional self-assessment rated.2)In this paper,VR equipment,EEG acquisition equipment,Android mobile devices together constitute a complete experimental data acquisition system and emotional recognition system.The SAM experiment program of VR platform is designed,and the SQLite database is constructed and the real-time monitoring and display system of Android mobile terminal is built.3)This paper designs and realizes the emotion recognition algorithm based on time-frequency feature fusion of EEG.As the user in the VR scene inevitably shaking his head into the artifact signal,this article uses wavelet transform and independent component analysis method as far as possible to remove the brain in the head movement,eye movement and other artifacts;In order to make the mood,this paper uses the emotional brain area theory to collect 10-channels effective EEG signal,extract the EEG signal frequency domain and frequency domain characteristics,through the typical component analysis algorithm(EEG),the EEG signal can be used to reduce the EEG signal acquisition,The time-frequency characteristics of the fusion and dimension reduction,screening out important features.With the SVM as the emotion recognition classifier,the 5-cross-validation is used.The average recognition rate of the emotion recognition system is 84.43%.