脑电信号情绪识别研究综述

Overview of Emotion Recognition Based on EEG Signals

情绪是一种内部的主观体验,是人们对外界因素或自身刺激产生的生理及心理反应,情绪识别在医疗、教育、刑事审讯等领域具有重要意义。脑电信号具有不易伪装、实时差异性等特点,使其在情绪识别领域具有显著优势。从传统的脑电信号特征提取、信号分类方法、基于深度学习的脑电信号识别方法,以及各种算法的识别率与性能对比等方面,梳理脑电信号识别发展历程。对脑电信号常用数据集进行对比并分析各数据集的优势与不足,总结出生成对抗网络(GAN)等用于脑电信号数据增强的技术与方法。最后对脑电信号情绪识别存在的问题和未来研究方向进行了讨论和展望。

Emotion is an internal subjective experience,which is the physiological and psychological response of people to external factors or their own stimulation.Emotion recognition is of great significance in medical treatment,education,criminal interrogation and other fields.EEG signal is not easy to camouflage,real-time differences and has other characteristics,which make it show significant advantages in the field of emotion recognition.Introduces the development of EEG signal recognition from five aspects:traditional EEG feature extraction,signal classification method and EEG signal recognition method based on deep learning,and analyzes and compares the recognition rate and performance of various algorithms.Then,the common EEG data sets are compared,and the advantages and disadvantages of each data set are analyzed.Aiming at these problems,the generation countermeasure network(GAN)and other technologies and methods for EEG data enhancement are summarized and introduced.Finally,the problems and future research directions of emotion recognition of EEG are discussed and prospected.