Research And Implementation Of VR Scene Construction Based On EEG

In recent years,the research based on EEG and virtual reality(VR)has developed rapidly.Combining EEG and VR technology has been widely used in psychological intervention.It can treat psychological diseases such as acrophobia and anxiety disorder,and relieve psychological pressure,achieve personalized and intelligent treatment.There are two difficulties in the research of VR scene construction based on EEG: one is emotion recognition based on EEG;and the other is emotion rating of VR scene content.At present,there are commercial-grade EEG emotion recognition products,but there is a lack of research on the relationship between VR scene content and emotions.Therefore,this paper conducts VR scene construction research work based on EEG signals.The main work is as follows:(1)Research on the relationship between VR scene content and emotions.Based on the International Affective Picture System(IAPS),emotional materials are selected to construct multiple emotional VR scenes with different stimulus variables.Then,the Self-Assessment Manikin(SAM)and the prefrontal lobe EEG signal are used to analyze the emotional relationship of the subjects under the stimulation of different virtual objects in VR scene.The experimental results show that the emotional state of the subjects shows corresponding obvious changes in the face of different virtual stimuli in VR scene,and the subjective assessment of SAM of the subjects is consistent with the emotional changes reflected by the EEG signal,so as to establish the corresponding relationship between part of the VR scene content and emotions.(2)Emotion prediction of VR scenes.In order to expand the corresponding relationship between VR scene content and emotion more quickly,determine what kind of emotion will be evoked by stimulus content added in VR scene to avoid negative effects,and provide reference for what kind of emotional state will be evoked by the VR scene that constructed based on low-dimensional emotional materials,this paper uses BP neural network and CNN to predict emotions of VR scene through experimental SAM data,and the mean absolute percentage error of the CNN prediction result is as low as 2.5%,which proves that the prediction experiment has a certain feasibility and application value.(3)EEG adaptive VR scene system design.Toward mental health applications,this paper uses HTC vive helmet and Brain Link EEG devices to build an emotional adaptive VR system based on EEG signal,which aims to relieve the psychological pressure in the study and work.