基于SDAE与RELM的EEG情感识别方法

doi:10.19678/j.issn.1000-3428.0059056

摘要/Abstract

摘要： 针对情感识别中堆叠式自动编码器存在反向传播方法收敛速度慢和容易陷入局部最优的问题，提出一种基于堆叠式降噪自动编码器（SDAE）和正则化极限学习机（RELM）的情感状态识别方法。从脑电信号的时域、频域和时频域中提取表征情感状态的初始特征，使用SDAE进行无监督特征学习，提取初始特征的高层抽象表示。在网络的回归层，使用RELM进行情感分类。在DEAP数据集上的实验结果表明，与SDAE以及DT、KNN等传统基于机器学习的方法相比，该方法在实时性、准确性和泛化性能等方面均有明显提升。

关键词: 情感识别, 脑电信号, 情感特征, 堆叠式降噪自动编码器, 正则化极限学习机

Abstract: The Stacked Auto-Encoders(SAE) used by the existing emotion recognition methods are limited by the low convergence speed at the back propagation stage, and tend to fall into local optimality.To address the problem, an emotion recognition method is proposed based on Stacked Denoising Auto-Encoder(SDAE) and Regularized Extreme Learning Machine(RELM).The method first requires the extraction of the initial features that characterize the emotional state from the time domain, frequency domain, and time-frequency domain of the Electroencephalogram(EEG) signals.Then SDAE is used for unsupervised feature learning to extract high-level abstract representations of the initial features.In the regression layer of the network, RELM is used for emotion classification.The experimental results on the DEAP data set show that the proposed method displays significant improvements in real-time performance, accuracy and generalization performance compared with SDAE, DT, KNN and other machine learning-based methods.

Key words: emotion recognition, Electroencephalogram(EEG) signals, emotional features, Stacked Denoising Auto-Encoder(SDAE), Regularized Extreme Learning Machine(RELM)

中图分类号:

TP18

连卫芳, 晁浩, 刘永利. 基于SDAE与RELM的EEG情感识别方法[J]. 计算机工程, 2021, 47(9): 75-83.

LIAN Weifang, CHAO Hao, LIU Yongli. EEG Emotion Recognition Method Based on SDAE and RELM[J]. Computer Engineering, 2021, 47(9): 75-83.

http://www.ecice06.com/CN/Y2021/V47/I9/75

图/表 15

20210917190143

20210917190146

20210917190150

20210917190154

20210917190158

20210917190201

20210917190205

20210917190208

20210917190213

20210917190217

20210917190221

20210917190225

20210917190228

20210917190232

20210917190236

参考文献

[1] ZHANG Y D, YANG Z J, LU H M, et al.Facial emotion recognition based on biorthogonal wavelet entropy, fuzzy support vector machine, and stratified cross validation[J].IEEE Access, 2017, 4:8375-8385.
[2] HUANG Z, DONG M, MAO Q, et al.Speech emotion recognition using CNN[C]//Proceedings of the 22nd ACM International Conference on Multimedia.New York, USA:ACM Press, 2014:801-804.
[3] 蒋静芳, 曾颖, 林志敏, 等.基于脑电信号的情绪评估研究综述[J].信息工程大学学报, 2016(6):686-693. JIANG J F, ZENG Y, LIN Z M, et al.Review of emotion evaluation based on EEG signals[J].Journal of Information Engineering University, 2016(6):686-693.(in Chinese)
[4] LIU Y, SOURINA O.Real-time fractal-based valence level recognition from EEG[M].Berlin, Germany:Springer, 2013.
[5] 高宁化, 王姮, 冯兴华.基于动态模糊决策树的心电信号分类方法[J].计算机工程, 2020, 46(1):80-86. GAO N H, WANG H, FENG X H.Classification method of electrocardiogram signals based on dynamic fuzzy decision tree[J].Computer Engineering, 2020, 46(1):80-86.(in Chinese)
[6] CHEN X, XU X, LIU A, et al.The use of multivariate EMD and CCA for denoising muscle artifacts from few-channel EEG recordings[J].IEEE Transactions on Instrumentation and Measurement, 2018, 67(2):359-370.
[7] 陈田, 陈占刚, 袁晓辉, 等.基于脑电信号瞬时能量的情感识别方法[J].计算机工程, 2019, 45(4):196-204. CHEN T, CHEN Z G, YUAN X H, et al.Emotion recognition method based on instantaneous energy of Electroencephalography[J].Computer Engineering, 2019, 45(4):196-204.(in Chinese)
[8] STOBER S, STERNIN A, OWEN A M, et al.Deep feature learning for EEG recordings[J].Computer Science, 2015, 165:23-31.
[9] YIN Z, ZHAO M Y, WANG Y X, et al.Recognition of emotions using multimodal physiological signals and an ensemble deep learning model[J].Computer Methods and Programs in Biomedicine, 2017, 140:93-110.
[10] LI K, LI X Y, ZHANG Y, et al.Affective state recognition from EEG with deep belief networks[C]//Proceedings of 2013 IEEE International Conference on Bioinformatics and Biomedicine.Washington D.C., USA:IEEE Press, 2013:305-310.
[11] JIA X W, LI K, LI X Y, et al.A novel semi-supervised deep learning framework for affective state recognition on EEG signals[C]//Proceedings of the 14th International Conference on Bioinformatics and Bioengineering.Washington D.C., USA:IEEE Press, 2014:30-37.
[12] PANDEY P, SEEJA K.Subject independent emotion recognition from EEG using VMD and deep learning[EB/OL].[2020-06-03].https://doi.org//10.1016/j.jksuci.2019.11.003.
[13] ZENG H, WU Z H, ZHANG J M, et al.EEG emotion classification using an improved SincNet-based deep learning model[J].Brain Sciences, 2019, 9(11):326-340.
[14] KOELSTRA S, MUHL C, SOLEYMANI M, et al.DEAP:a database for emotion analysis; using physiological signals[J].IEEE Transactions on Affective Computing, 2012, 3(1):18-31.
[15] HUANG G B, ZHU Q Y, SIEW C K.Extreme learning machine:a new learning scheme of feed forward neural networks[C]//Proceedings of IEEE International Joint Conference on Neural Networks.Washington D.C., USA:IEEE Press, 2004:985-990.
[16] FLETCHER R.Practical methods of optimization:constrained optimization[M].Washington D.C., USA:IEEE Press, 2013.
[17] WEN Z Y, XU R F, DU J C.A novel convolutional neural networks for emotion recognition based on EEG signal[C]//Proceedings of International Conference on Security.Washington D.C., USA:IEEE Press, 2018:672-677.
[18] ZHUANG N, ZENG Y, LI T, et al.Emotion recognition from EEG signals using multidimensional information in EMD domain[EB/OL].[2020-06-03].https://downloads.hindawi.com/journals/bmri/2017/8317357.pdf.
[19] XING X F, LI Z Q, XU T Y, et al.SAE+LSTM:a new framework for emotion recognition from multi-channel EEG[EB/OL].[2020-06-03].https://doi.org/10.3389/fnbot.2019.00037.
[20] CHEN M, HAN J W, GUO L, et al.Identifying valence and arousal levels via connectivity between EEG channels[C]//Proceedings of International Conference on Affective Computing & Intelligent Interaction.Washington D.C., USA:IEEE Press, 2015:63-69.
[21] ZHANG P, LI X, HOU Y X, et al.EEG based emotion identification using unsupervised deep feature learning[C]//Proceedings of SIGIR 2015 Workshop on Neuro-Physiological Methods in IR Research.Washington D.C., USA:IEEE Press, 2015:1-12.
[22] LI X, SONG D W, ZHANG P, et al.Emotion recognition from multi-channel EEG data through convolutional recurrent neural network[C]//Proceedings of IEEE International Conference on Bioinformatics and Biomedicine.Washington D.C., USA:IEEE Press, 2017:352-359.
[23] DAIMI S, SAHA G.Classsification of emotions induced by music videos and correlation with participants'rating[J].Expert Systems with Applications, 2014, 41(13):6057-6065.

选择文件类型/文献管理软件名称

选择包含的内容