南京师范大学学报（工程技术版）2025，Vol.25Issue(2)：102-108,7.DOI:10.3969/j.issn.1672-1292.2025.02.010
基于脑电信号的情感识别SimAM-CRNN方法
SimAM-CRNN Model for Emotion Recognition Based on EEG Signals
摘要
Abstract
Emotion recognition in brain-computer interfaces identifies emotional states by analyzing electroencephalogram(EEG)signals.This process is vital for applications such as disease monitoring and workload assessment.Althrough deep learning methods have advanced EEG-based emotion recognition,challenges remain in fully exploring multi-domain information from these signals.To address these challenges,a four-dimensional attention neural network,SimAM-CRNN,is proposed.This method transforms the original EEG signals into four-dimensional representations encompassing spectra,space,and time.It integrates SimAM with convolutional neural networks(CNN)and long short-term memory networks(LSTM)for more comprehensive feature extraction.The proposed method is validated by using the public DEAP dataset,achieving a valence accuracy of 95.29%and an arousal accuracy of 95.62%.关键词
脑电信号/情感识别/深度学习/注意力机制Key words
EEG signals/emotion recognition/deep learning/attention mechanism分类
信息技术与安全科学引用本文复制引用
梁起鹏,耿晓中,户唯新,陈成,于萍..基于脑电信号的情感识别SimAM-CRNN方法[J].南京师范大学学报（工程技术版）,2025,25(2):102-108,7.基金项目
吉林省科技发展计划项目(20240404059ZP). （20240404059ZP）
