SpecMNet: Spectrum mend network for monaural speech enhancement

doi:10.1016/j.apacoust.2022.108792

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

	SpecMNet: Spectrum mend network for monaural speech enhancement
	Fan, Cunhang 5; Zhang, Hongmei 5; Yi, Jiangyan 3; Lv, Zhao 3,5; Tao, Jianhua 2,3; Li, Taihao 4; Pei, Guanxiong 4; Wu, Xiaopei 5; Li, Sheng 1
刊名	APPLIED ACOUSTICS
	2022-06-15
卷号	194 页码:9
关键词	Monaural speech enhancement Speech distortion Spectrum mend network SI-SNR BLSTM
ISSN号	0003-682X
DOI	10.1016/j.apacoust.2022.108792
通讯作者	Yi, Jiangyan(jiangyan.yi@nlpr.ia.ac.cn) ; Lv, Zhao(kjlz@ahu.edu.cn) ; Tao, Jianhua(jhtao@nlpr.ia.ac.cn)
英文摘要	Speech enhancement methods usually suffer from speech distortion problem, which leads to the enhanced speech losing so much significant speech information. This damages the speech quality and intelligibility. In order to address this issue, we propose a spectrum mend network (SpecMNet) for monaural speech enhancement. The proposed SpecMNet aims to retrieve the lost information by mending the weighted enhanced spectrum with weighted original spectrum. More specifically, the proposed algorithm consists of pre-enhancement network and the mend network. The main task of preenhancement network is to acquire the pre-enhanced spectrum so that it can remove the most of the noise signals. Because of the speech distortion problem, it loses a great deal of speech components. While the original spectrum has no speech information lost. Therefore, we utilize the original spectrum to mend the pre-enhanced spectrum by adding these two weighted spectrums so that the lost speech information can be retrieved. Then the mend network is used to predict mend weights for these two spectrums. Finally, the mended spectrum is used as the enhanced output. Our experiments are conducted on the TIMIT + (100 Nonspeech Sounds and NOISEX-92) datasets. Experimental results demonstrate that our proposed SpecMNet approach is effective to alleviate the speech distortion problem. (c) 2022 Elsevier Ltd. All rights reserved.
资助项目	National Key Research and Development Program of China[2021ZD0201502] ; National Natural Science Foundation of China (NSFC)[61972437] ; Open Research Projects of Zhejiang Lab[2021KH0AB06] ; Open Projects Program of National Laboratory of Pattern Recognition[202200014]
WOS关键词	NEURAL-NETWORK ; NOISE
WOS研究方向	Acoustics
语种	英语
出版者	ELSEVIER SCI LTD
WOS记录号	WOS:000798344800011
资助机构	National Key Research and Development Program of China ; National Natural Science Foundation of China (NSFC) ; Open Research Projects of Zhejiang Lab ; Open Projects Program of National Laboratory of Pattern Recognition
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/49551]
专题	模式识别国家重点实验室_智能交互
通讯作者	Yi, Jiangyan; Lv, Zhao; Tao, Jianhua
作者单位	1.Natl Inst Informat & Commun Technol, Kyoto, Japan 2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 3.Chinese Acad Sci, Inst Automat, NLPR, Beijing 100190, Peoples R China 4.Artificial Intelligence Res Inst, Zhejiang Lab, Hangzhou 311121, Peoples R China 5.Anhui Univ, Sch Comp Sci & Technol, Anhui Prov Key Lab Multimodal Cognit Computat, Hefei 230601, Peoples R China
推荐引用方式 GB/T 7714	Fan, Cunhang,Zhang, Hongmei,Yi, Jiangyan,et al. SpecMNet: Spectrum mend network for monaural speech enhancement[J]. APPLIED ACOUSTICS,2022,194:9.
APA	Fan, Cunhang.,Zhang, Hongmei.,Yi, Jiangyan.,Lv, Zhao.,Tao, Jianhua.,...&Li, Sheng.(2022).SpecMNet: Spectrum mend network for monaural speech enhancement.APPLIED ACOUSTICS,194,9.
MLA	Fan, Cunhang,et al."SpecMNet: Spectrum mend network for monaural speech enhancement".APPLIED ACOUSTICS 194(2022):9.