Electroencephalogram emotion recognition method based on graph convolutional neural network

Abstract

The invention belongs to the technical field of signal processing, and particularly relates to an electroencephalogram emotion recognition method based on a graph convolutional neural network. The electroencephalogram emotion recognition method provided by the invention comprises the following steps: firstly, constructing an emotion-related function connection mode according to the correlation between electroencephalogram channel signals; then modeling each extracted single-channel electroencephalogram feature into a graph signal based on a function connection mode structure, and establishinga graph convolutional neural network model; inputting a training set graph signal to carry out iterative optimization training on the model, and optimizing network parameters by adopting a back propagation algorithm and a cross entropy loss function; and finally, inputting a test set graph signal into the trained graph convolutional neural network, and performing classification recognition on thetiter, wake-up and dominance three-dimensional emotion model. According to the electroencephalogram emotion recognition method, the spatial and functional relationship of the electroencephalogram datais restored, so that electroencephalogram emotion features with higher discriminability can be extracted; and the channel electrode most relevant to emotion can be selected according to the channel correlation, and processing time and hardware cost are reduced while the number of channels is reduced.

Description

technical field [0001] The invention belongs to the technical field of signal processing, and more specifically relates to an EEG emotion recognition method based on a graph convolutional neural network. Background technique [0002] Among the many physiological signals, because EEG is directly obtained from the brain signal, it can directly reflect the state of brain activity, and it is easy to extract, with high time resolution and strong real-time performance. Therefore, it has received extensive attention and developed rapidly in emotion recognition research. The study of EEG signals under different emotional states has gradually formed an emerging interdisciplinary research, which not only includes computer science and artificial intelligence, but also involves neurocognitive science. Effectively interpreting the emotional state of individuals can solve various practical problems in work and life, such as monitoring people's mental health in work and life, understanding...

AI Technical Summary

Problems solved by technology

Among them, convolutional neural networks have been widely used in spatial continuous data processing such as computer vision and natural language processing. It is worth noting that convolutional neural networks cannot handle spatially discrete data, such as biomolecules, social networks, and brain networks.

Since the number of adjacent vertices of each vertex in the topological graph may be different, it is impossible to use a convolution kernel of the same size for convolution operation, so the traditional discrete convolution cannot maintain translation on non-Euclidean data. transsexual

Images