Multi-Level Attention-Based Categorical Emotion Recognition Using Modulation-Filtered Cochleagram

doi:10.3390/app13116749

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 智能交互

	Multi-Level Attention-Based Categorical Emotion Recognition Using Modulation-Filtered Cochleagram
	Peng, Zhichao 4; He, Wenhua 4; Li, Yongwei 1; Du, Yegang 5; Dang, Jianwu 2,3
刊名	APPLIED SCIENCES-BASEL
	2023-06-01
卷号	13 期号:11 页码:16
关键词	categorical emotion recognition auditory signal processing modulation-filtered cochleagram multi-level attention
DOI	10.3390/app13116749
通讯作者	Peng, Zhichao(zcpeng@tju.edu.cn) ; Dang, Jianwu(jdang@jaist.ac.jp)
英文摘要	Speech emotion recognition is a critical component for achieving natural human-robot interaction. The modulation-filtered cochleagram is a feature based on auditory modulation perception, which contains multi-dimensional spectral-temporal modulation representation. In this study, we propose an emotion recognition framework that utilizes a multi-level attention network to extract high-level emotional feature representations from the modulation-filtered cochleagram. Our approach utilizes channel-level attention and spatial-level attention modules to generate emotional saliency maps of channel and spatial feature representations, capturing significant emotional channel and feature space from the 3D convolution feature maps, respectively. Furthermore, we employ a temporal-level attention module to capture significant emotional regions from the concatenated feature sequence of the emotional saliency maps. Our experiments on the Interactive Emotional Dyadic Motion Capture (IEMOCAP) dataset demonstrate that the modulation-filtered cochleagram significantly improves the prediction performance of categorical emotion compared to other evaluated features. Moreover, our emotion recognition framework achieves comparable unweighted accuracy of 71% in categorical emotion recognition by comparing with several existing approaches. In summary, our study demonstrates the effectiveness of the modulation-filtered cochleagram in speech emotion recognition, and our proposed multi-level attention framework provides a promising direction for future research in this field.
资助项目	Hunan Provincial Natural Science Foundation of China[2021JJ30379] ; Youth Fund of the National Natural Science Foundation of China[62201571]
WOS关键词	FEATURES
WOS研究方向	Chemistry ; Engineering ; Materials Science ; Physics
语种	英语
出版者	MDPI
WOS记录号	WOS:001003438300001
资助机构	Hunan Provincial Natural Science Foundation of China ; Youth Fund of the National Natural Science Foundation of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/53498]
专题	模式识别国家重点实验室_智能交互
通讯作者	Peng, Zhichao; Dang, Jianwu
作者单位	1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100045, Peoples R China 2.Tianjin Univ, Coll Intelligence & Comp, Tianjin 300350, Peoples R China 3.Pengcheng Lab, Shenzhen 518055, Peoples R China 4.Hunan Univ Humanities Sci & Technol, Informat Sch, Loudi 417000, Peoples R China 5.Waseda Univ, Future Robot Org, Tokyo 1698050, Japan
推荐引用方式 GB/T 7714	Peng, Zhichao,He, Wenhua,Li, Yongwei,et al. Multi-Level Attention-Based Categorical Emotion Recognition Using Modulation-Filtered Cochleagram[J]. APPLIED SCIENCES-BASEL,2023,13(11):16.
APA	Peng, Zhichao,He, Wenhua,Li, Yongwei,Du, Yegang,&Dang, Jianwu.(2023).Multi-Level Attention-Based Categorical Emotion Recognition Using Modulation-Filtered Cochleagram.APPLIED SCIENCES-BASEL,13(11),16.
MLA	Peng, Zhichao,et al."Multi-Level Attention-Based Categorical Emotion Recognition Using Modulation-Filtered Cochleagram".APPLIED SCIENCES-BASEL 13.11(2023):16.