Emotion recognition method and system fusing prior and automatic electroencephalogram characteristics

The invention discloses an emotion recognition method and system fusing prior and automatic electroencephalogram characteristics, and the method comprises the steps: firstly extracting different frequencies of difference entropy characteristics related to emotion from an original electroencephalogram signal, then mapping the difference entropy characteristics into a difference entropy matrix, splicing the difference entropy matrix into a space-frequency characteristic matrix, and taking the space-frequency characteristic matrix as a mathematically-driven prior emotion characteristic; automatically extracting time-frequency domain information in the original electroencephalogram signals by scaling a convolutional layer to serve as automatic electroencephalogram emotion features driven by data; transforming, splicing and fusing the priori emotion features and the automatic electroencephalogram emotion features, extracting high-order semantic features of the features, and finally sending the high-order semantic