A Comparative Study of Machine Learning Techniques for Emotion Recognition from Peripheral Physiological Signals

被引:0
作者
Vijayakumar, Sowmya [1 ]
Flynn, Ronan [1 ]
Murray, Niall [1 ]
机构
[1] Athlone Inst Technol, Dept Comp & Software Engn, Athlone, Co Westmeath, Ireland
来源
2020 31ST IRISH SIGNALS AND SYSTEMS CONFERENCE (ISSC) | 2020年
关键词
peripheral physiological signals; emotion recognition; machine learning; wearables; classification;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Recent developments in wearable technology have led to increased research interest in using peripheral physiological signals for emotion recognition. The non-invasive nature of peripheral physiological signal measurement via wearables enables ecologically valid long-term monitoring. These peripheral signal measurements can be used in real-time in many ways including health and emotion classification. This paper investigates the utility of peripheral physiological signals for emotion recognition using the publicly available DEAP database. Using this database (which contains electroencephalogram (EEG) signals and peripheral signals), this paper compares eight machine learning models in the classification of valence and arousal emotion dimensions. These were applied to the peripheral physiological signals only. These models operate on three groupings of the peripheral data: (i) the raw peripheral physiological signals; (ii) individual feature sets extracted from each peripheral signal; and (iii) a fusion data set made of the combined features from the individual peripheral signals. The results indicate that support vector machine, linear discriminant analysis and logistic regression give the best recognition results on all three data groups considered. The feature fusion data set, which is made up by fusing all the features from the peripheral signals, gives the best recognition accuracy on both valence and arousal dimensions. In addition, subject dependency for emotion classification from peripheral signals is examined and significant individual variability is observed. The recognition rate varies between each participant from 10% to 87.5%.
引用
收藏
页码:80 / 85
页数:6
相关论文
共 18 条
[1]   Emotion Recognition from Multimodal Physiological Signals for Emotion Aware Healthcare Systems [J].
Ayata, Deger ;
Yaslan, Yusuf ;
Kamasak, Mustafa E. .
JOURNAL OF MEDICAL AND BIOLOGICAL ENGINEERING, 2020, 40 (02) :149-157
[2]  
Ayata D, 2015, 2016 MEDICAL TECHNOLOGIES NATIONAL CONFERENCE (TIPTEKNO)
[3]  
Chatterjee Abhijit, 2016, 2016 IEEE 21st International Mixed-Signal Testing Workshop (IMSTW). Proceedings, P1, DOI 10.1109/IMS3TW.2016.7524222
[4]   Subject-independent emotion recognition based on physiological signals: a three-stage decision method [J].
Chen, Jing ;
Hue, Bin ;
Wang, Yue ;
Moore, Philip ;
Dail, Yongqiang ;
Feng, Lei ;
Dingo, Zhijie .
BMC MEDICAL INFORMATICS AND DECISION MAKING, 2017, 17
[5]  
Chen J, 2015, IEEE INT C BIOINFORM, P395, DOI 10.1109/BIBM.2015.7359713
[6]  
Chen SY, 2016, INT CONF ACOUST SPEE, P2827, DOI 10.1109/ICASSP.2016.7472193
[7]   Cancer incidence and mortality in China, 2014 [J].
Chen, Wanqing ;
Sun, Kexin ;
Zheng, Rongshou ;
Zeng, Hongmei ;
Zhang, Siwei ;
Xia, Changfa ;
Yang, Zhixun ;
Li, He ;
Zou, Xiaonong ;
He, Jie .
CHINESE JOURNAL OF CANCER RESEARCH, 2018, 30 (01) :1-12
[8]  
Godin C., PHYCS 2015
[9]  
He L., 2015, P AVECACM MULT, P73
[10]   Deep learning analysis of mobile physiological, environmental and location sensor data for emotion detection [J].
Kanjo, Eiman ;
Younis, Eman M. G. ;
Ang, Chee Siang .
INFORMATION FUSION, 2019, 49 :46-56