Multi-Channel EEG Signal Processing in Emotion Recognition by Segmentation and Optimisation Using Machine Learning Algorithms

Abstract

Human decision-making, planning, thinking, and other mental processes are all greatly influenced by emotions, which are an essential part of everyday life. They are therefore regarded as an important component of human interactions. Human emotions are recognised from different sources, combining speech, body language, facial expressions and behaviour (position and gesture). The brain's electroencephalogram (EEG) impulses provide us with a more varied understanding of emotional states that people might not be able to articulate. This research proposes novel technique in emotion recognition based on segmentation with optimisation using multi-channel based EEG with Machine Learning (ML) model. Emotion recognition utilizingmulti-channel EEG signals with ML involve analyzing brain activity captured by two EEG channels to identify emotional states, leveraging ML algorithms for classification. EEG signals collects the inputs are processed for the noise removal process. Then this signal has been segmented utilizing multi-channel graph cut attention convolutional neural networks. From this segmented the EEG signal has been pipelined based on discrete residual transform with artificial grey wolf swarm optimization. The experimental Analysis was done in terms of finding accuracy, mean average precision, recall, F-1 score, AUC. The proposed model achieved detection accuracy. is 97%, mean average precision is 95 %, recall is 93%, F-1 score is 91%, AUC is 90%.