Emotion Recognition Of The Deaf Based On Double Threshold Brain Network

Emotions are the adaptive psychological and physiological expressions produced when the people is stimulated by the outside.Due to the lack of hearing ability,the deaf people may have poor social adaptation,emotional expression disorders and emotional deviations in the process of growing up.A new method was proposed for emotion recognition based on EEG of the deaf.The main study contents are as follows:(1)The experimental paradigm is designed for inducing the positive,neutral,and negative emotions of the deaf by movies,and the EEG signals of the deaf were collected.The EEG dataset containing 15 deaf people is established.They performed emotion annotation according to the SAM system after each movie clip is played,then the signals are screened according to the annotation to ensure the reliability of the experiment.(2)The framework based on dual-threshold brain network is designed for emotion recognition of the deaf people.First,the phase-locked value is used to calculate the coupling relationship between nodes.The double-threshold method is used to construct the brain network to avoid the problem that the single-threshold value is too high cause the important connection is discarded;too low cause unnecessary connections is increased.Then,the global features and local features of the brain network are extracted,and the stacking ensemble learning model is used for classification.Finally,the extracted brain network features are fused with power spectral density and differential entropy features respectively,which effectively improves the accuracy.(3)A sample-normalized statistical method for discriminative brain regions of key emotions is proposed,which is used to select the key electrodes related to the emotions of the deaf.First,by drawing a brain topology map of the energy characteristic value of the deaf,the activity of each brain area in different emotional states is analyzed.Then,6 key electrode selection modes were designed,namely 4 channels,6 channels,9 channels,12-d channels,12-e channels,and 23 channels for emotion recognition.Finally,the extracted six patterns of EEG features are classified to obtain the accuracy of emotion recognition corresponding to different channels,and then the relationship between brain regions and emotions of the deaf is analyzed.