Design Of Emotion Recognition And Negative Emotion Assisted Intervention System Based On Eeg Feature Fusion

With the in-depth development of human-computer interaction research,emotion recognition has been more and more widely used in medical care,education,service and other fields.Emotion is closely related to cognition,and a good emotional state will have a positive feedback effect on attention and memory.Therefore,the effective recognition of emotional state and auxiliary intervention of negative emotions have attracted researchers’ attention.Because of its objectivity,authenticity and high temporal resolution,EEG signals have shown great potential in emotion recognition.Emotion recognition and auxiliary intervention based on EEG signals are developing continuously and have become a research hotspot.In this paper,a negative emotion assist intervention system based on multi-feature fusion of EEG signals is proposed for negative emotion assist intervention.The system integrates the time-frequency characteristics and nonlinear characteristics of EEG signals to realize emotion recognition,and uses the recognition result as the feedback index to realize the auxiliary feedback regulation of negative emotions.Will DEAP of international standards for public emotion recognition database and the combination of eeg signals,multiangle extraction and analyzed the characteristics of the emotional brain electrical signal parameters,including: the standard deviation,second-order difference time domain average characteristics of the beta power frequency domain partial lateral asymmetry index,spectrum characteristics,entropy and LZ Complexity and nonlinear characteristic parameters of sample(Lempel-Ziv Complexity,LZC)characteristics.Considering the realtime performance and feature redundancy of the system,feature selection is realized based on Relief feature algorithm.Based on support vector machine,positive,negative and neutral emotions were classified,and the recognition accuracy reached 92.3%.In the negative emotion assisted intervention system,Emotiv EPOC + was used as the EEG signal acquisition device,the system platform was built based on Open Vib,and the expression of music feedback information was completed by Unity software.The system is a comprehensive platform which can realize EEG signal acquisition,emotion recognition and negative emotion auxiliary regulation based on the recognition result.Based on this system,emotion recognition and negative emotion assisted feedback regulation tests were designed and completed for subthreshold depressed subjects.The results showed that the scores of SDS and PANAS were significantly improved after the test(P <0.05).In terms of EEG signal characteristics,the relative power values of α wave and βwave and sample entropy characteristics of subjects were significantly increased(P <0.05).Except for F3,the second-order difference characteristic values of electrodes in the left brain region showed an upward trend,while the second-order difference characteristic values of P8,FC6 and F4 electrodes in the right brain region showed a downward trend.The effectiveness of the negative emotion assisted intervention system designed in this paper is verified.