In response to the problems of low efficiency and high sensitivity of traditional emotion inducing materials,this study investigates the use of virtual reality technology to establish an emotion inducing system,and uses a dual supervised neural network model for emotion recognition,building an emotion classification technique using a classification algorithm and regression dual loss function constraints.On this basis,a multi perspective mechanism is introduced to achieve automatic annotation and improve the ac-curacy and efficiency of emotion recognition.The results showed that the time proportion of positive emotional confidence in fairy tale park scenes reached 85.75％,the time proportion of negative emotional confidence interval in deep cave scenes reached 95％,and the time proportion of neutral emotional confidence in scenic countryside and seaside sunset scenes was 90.68％and 88.76％,respec-tively.This confirms the effectiveness of multi perspective mechanisms for mental health monitoring and virtual reality integration technology.
维普
万方数据