Author Login Editor-in-Chief Peer Review Editor Work Office Work

Computer Engineering ›› 2022, Vol. 48 ›› Issue (1): 69-74. doi: 10.19678/j.issn.1000-3428.0060182

• Artificial Intelligence and Pattern Recognition • Previous Articles     Next Articles

Emotion Recognition Using Causal Network Based on Multi-Channel EEG Signal

WANG Bin1, WANG Zhongmin1,2, ZHANG Rong1,2   

  1. 1. School of Computer Science and Technology, Xi'an University of Posts and Telecommunications, Xi'an 710121, China;
    2. Shaanxi Key Laboratory of Network Data Analysis and Intelligent Processing, Xi'an 710121, China
  • Received:2020-12-03 Revised:2021-01-11 Published:2021-01-14

基于多通道脑电信号的因果网络情绪识别

王斌1, 王忠民1,2, 张荣1,2   

  1. 1. 西安邮电大学 计算机学院, 西安 710121;
    2. 陕西省网络数据分析与智能处理重点实验室, 西安 710121
  • 作者简介:王斌(1992-),男,硕士研究生,主研方向为深度学习、情感计算;王忠民,教授、博士;张荣,讲师、硕士。
  • 基金资助:
    国家自然科学基金(61373116);陕西省科技厅工业领域一般项目(2018GY-013);陕西省教育厅专项科学研究计划项目(18JK0697);咸阳市科学技术研究计划项目(2017k01-25-2)。

Abstract: Emotion is produced by multiple channels in the brain.As a mainstream method of emotion recognition, Granger causality tends to ignore the influence of other channels when calculating the causality between any two channels.For multi-channel EEG signal, a emotion recognition method using causal network based on Conditional Granger Causality test(CGC) is proposed.The CGC algorithm is used to calculate the causal relationship of the whole brain channel under different emotions, and then the causal network is constructed.By analyzing the in degree, out degree and intermediate topological attributes of each channel, the key channel is found, and the simplified causal network is obtained for emotion recognition.The causal connection between nodes is input into SVM and KNN classifiers for classification training.The experimental results show that the recognition rates of the simplified network are 75.3% and 78.4% respectively, which verifies the effectiveness of the proposed method.

Key words: Conditional Granger Causal test(CGC), EEG signal, causal network, key channel, emotion recognition

摘要: 情绪是由大脑内多个通道共同作用产生的,格兰杰因果检验作为情绪识别的主流方法,在计算任意2个通道之间的因果关系时容易忽略其他通道的影响。面向多通道脑电信号,提出一种基于条件格兰杰因果检验(CGC)的因果网络情绪识别方法。利用CGC算法计算不同情绪下大脑全通道的因果关系,据此构建因果网络,并通过分析各通道的入/出度和介数拓扑属性找到关键通道,得到简化的因果网络进行情绪识别。将节点之间的因果连接关系作为特征分别输入SVM和KNN分类器进行分类训练,实验结果表明,简化网络的识别率分别为75.3%和78.4%,验证了所提方法的有效性。

关键词: 条件格兰杰因果检验, 脑电信号, 因果网络, 关键通道, 情绪识别

CLC Number: