Automatic Emotion Recognition from EEG Signals Using a Combination of Type-2 Fuzzy and Deep Convolutional Networks

Abstract

Emotions are an inextricably linked component of human life. Automatic emotion recognition can be widely used in brain–computer interfaces. This study presents a new model for automatic emotion recognition from electroencephalography signals based on a combination of deep learning and fuzzy networks, which can recognize two different emotions: positive, and negative. To accomplish this, a standard database based on musical stimulation using EEG signals was compiled. Then, to deal with the phenomenon of overfitting, generative adversarial networks were used to augment the data. The generative adversarial network output is fed into the proposed model, which is based on improved deep convolutional networks with type-2 fuzzy activation functions. Finally, in two separate class, two positive and two negative emotions were classified. In the classification of the two classes, the proposed model achieved an accuracy of more than 98%. In addition, when compared to previous studies, the proposed model performed well and can be used in future brain–computer interface applications.