Electroencephalogram emotion recognition method based on generative adversarial network data enhancement

Abstract

The invention discloses an electroencephalogram emotion recognition method based on generative adversarial network data enhancement, which comprises the following steps of: firstly, determining the length of an electroencephalogram sample, the length of a slice in the sample and slice window shift, slicing the electroencephalogram sample, and decomposing the electroencephalogram sample into four different frequency bands; secondly, calculating electroencephalogram signal features on all channel frequency bands, splicing all the calculated features into one-dimensional feature vectors, inputting the one-dimensional feature vectors into a generative adversarial network for training, and after training is completed, the network generating a large number of high-quality electroencephalogram samples; rearranging the generated electroencephalogram feature vectors and the original electroencephalogram feature vectors according to electrode placement position information during electroencephalogram signal collection, splicing according to frequency bands, constructing a three-dimensional electroencephalogram feature map, sending the three-dimensional electroencephalogram feature map into a continuous convolutional neural network to be trained, and finally carrying out sentiment classification through a softmax classifier. According to the method, good efficiency and robustness can be kept under the situation that few samples are recognized in the electroencephalogram emotion.

Description

technical field [0001] The invention relates to the technical field of EEG emotion recognition, in particular to a method for EEG emotion recognition based on data enhancement of generative adversarial networks. Background technique [0002] Emotions are fundamental in human's daily life because they play an important role in human cognition, namely perception, rational decision-making, human interaction and human intelligence. With the development of artificial intelligence technology and deep learning, emotion recognition has broad prospects in human-computer interaction and clinical treatment, which has been widely concerned by researchers. Human emotions are recognized through speech, eye blinks, facial expressions and physiological signals. But the first three methods are unstable and susceptible to subjectivity. Subjects may deliberately hide their emotions and lead to identification errors. Physiological signals such as electrooculogram (EOG), electroencephalogram ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a method for EEG emotion recognition based on data enhancement of generative confrontation network, which solves the problems of insufficient samples and imbalance between sample classes in the training process of EEG emotion recognition model, and improves the accuracy and accuracy of emotion model recognition. Recall rate, this method can improve the feature description ability of different categories of emotion recognition model, so that the model has higher accuracy and better robustness

Images