基于脑电信号深度学习的情感分类

doi:10.3969/j.issn.0255-8297.2021.03.001

摘要/Abstract

摘要： 情感脑电研究作为人工智能高级阶段的重要任务，近年来受到越来越多的关注。情感脑电分类广泛应用于人机交互、医学研究等领域。该文以轻量级的卷积神经网络为核心，设计了情感脑电分类模型，以DEAP（dataset for emotion analysis using physiologicalsignals）提供的情感脑电图数据为基础，将其中的观看视频划分为唤醒度和愉悦度2个维度。为了获得频域信息，提取了theta、alpha、beta和gamma波段的功率谱密度特征进行评估，并将功率谱密度矩阵表示为二维灰度图像。然后将该图像输入到卷积神经网络训练分类模型并完成2个维度的分类任务。实验结果表明，与传统机器学习相比，卷积神经网络具有更好的分类效果，唤醒度分类准确率达到了82.33%，愉悦度分类准确率达到了75.46%。

关键词: 情感, 脑电, 深度学习, 卷积神经网络, 功率谱密度

Abstract: Electroencephalograph (EEG) research of emotion, as an important task in the advanced stage of artificial intelligence, has received more and more attention in recent years. Emotional EEG classification is widely used in human-computer interaction, medical research and other fields. This study presents the design of an EEG classification system on a lightweight convolutional neural network (CNN). DEAP (dataset for emotion analysis using physiological signals) provides EEG data of two kinds of emotion: arousal and valence. In order to obtain frequency domain information, the power spectral density features of theta, alpha, beta and gamma bands are extracted for evaluation, and each power spectral density matrix is expressed as a two-dimensional gray-scale image. The images were input into the convolutional neural network to train the classification model and complete the task of two classification. Experimental results show that compared with traditional machine learning, CNN has better classification effect. The accuracy of the two classification is 82.33% (Arousal) and 75.46% (Valence) respectively.

Key words: emotion, electroencephalograph (EEG), deep learning, convolutional neural network (CNN), power spectral density

中图分类号:

TP391

郝琰, 石慧宇, 霍首君, 韩丹, 曹锐. 基于脑电信号深度学习的情感分类[J]. 应用科学学报, 2021, 39(3): 347-346.

HAO Yan, SHI Huiyu, HUO Shoujun, HAN Dan, CAO Rui. Emotion Classification Based on EEG Deep Learning[J]. Journal of Applied Sciences, 2021, 39(3): 347-346.

参考文献

[1] 祝宇虹, 魏金海, 毛俊鑫, 等. 人工情感研究综述[J]. 江南大学学报(自然科学版), 2012, 11(4): 497-504. Zhu Y H, Wei J H, Mao J X, et al. Summary of artificial emotion[J]. Journal of Jiangnan University (Natural Science Edition), 2012, 11(4): 497-504. (in Chinese)
[2] 陆文娟. 基于脑电信号的情感识别研究[M]. 南京: 南京邮电大学, 2017.
[3] 张鑫, 范勇, 蒋田仔, 等. 基于脑电信号构建脑网络[C]//第一届全国神经动力学学术会议. 上海, 2012: 64-71.
[4] Pan C, Shi C, Mu H, et al. EEG-based emotion recognition using logistic regression with Gaussian kernel and Laplacian prior and investigation of critical frequency bands[J]. Applied Sciences, 2020, 10(5): 1619-1642.
[5] Liu S, Tong J J, Meng J Y. Study on an effective cross-stimulus emotion recognition model using EEGs based on feature selection and support vector machine[J]. International Journal of Machine Learning and Cybernetics, 2018, 9: 721-726.
[6] Hou H R, Zhang X N, Meng Q H. Odor-induced emotion recognition based on average frequency band division of EEG signals[J]. Journal of Neuroscience Methods, 2020, 334: 108599-108605.
[7] Jiang X M, Zhang J R, Chen F Q, et al. Emotion recognition based on J48 decision tree classifier and results analysis[J]. Computer Engineering and Design, 2017, 38(3): 761-767.
[8] Krizhevsky A, Sutskever G H. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84-90.
[9] Hermann K M, Kocisky T, Grefenstette E, et al. Teaching machines to read and comprehend[C]//The Neural Information Processing Systems Conference, Canada, Montreal, 2015: 1693-1701.
[10] Schirrmeister R T, Springenberg J T, Fiederer L D J, et al. Deep learning with convolutional neural networks for EEG decoding and visualization[J]. Human Brain Mapping, 2017, 38(11): 5391-5420.
[11] Mu Y, Han S H, Gelfand M J. The role of gamma interbrain synchrony in social coordination when humans face territorial threats[J]. Social Cognitive and Affective Neuroscience, 2017, 12(10): 1614-1623.
[12] 杨晓明, 晋玉剑, 李永红. 经典功率谱估计Welch法的MATLAB仿真分析[J]. 电子测试, 2011, 7(7): 101-104. Yang X M, Jin Y J, Li Y H. MATLAB simulation and analysis of the Welch method in the classical power spectrum estimation[J]. Electronic Test, 2011, 7(7): 101-104. (in Chinese)
[13] Lan M, James W, Thierry B, et al. Resting state EEG-based biometrics for individual identification using convolutional neural networks[C]//Conference Proceedings IEEE Engineering in Medicine and Biology Society, 2015: 2848-2851.
[14] Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[C]//The International Conference on Learning Representations, San Diego, USA, 2015: 1409-1556.
[15] Bashivan P, Rish I, Yeasin M, et al. Learning representations from EEG with deep recurrentconvolutional neural networks[C]//International Conference on Learning Representations. San Juan, Puerto Rico. 2016: 1511-1525.
[16] 赵文萍. 基于生理信号多模态情感识别研究[D]. 天津: 天津师范大学, 2019.
[17] 童中凯. 基于生理信号的情绪识别与睡眠分期的研究[D]. 河北: 燕山大学, 2019.
[18] 金雨鑫, 骆懿, 于洋. 基于深度深林的脑电情绪识别研究[J]. 软件导刊, 2019, 18(7): 53-55. Jin Y X, Luo Y, Yu Y. Research on EEG emotion recognition based on deep forest[J]. Software Guide, 2019, 18(7): 53-55. (in Chinese)
[19] Gupta V, Chopda M D, Ram B. Cross-subject emotion recognition using flexible analytic wavelet transform from EEG signals[J]. IEEE Sensors Journal, 2019, 19(6): 2266-2274.