Single-channel dereverberation for distant-talking speech recognition by combining denoising autoencoder and temporal structure normalization

被引：0

作者：

Ueda, Yuma ^{[1
]}

Wang, Longbiao ^{[2
]}

Kai, Atsuhiko ^{[1
]}

Xiao, Xiong ^{[3
]}

Chng, Eng Siong ^{[3
]}

Li, Haizhou ^{[4
]}

机构：

[1] Shizuoka Univ, Grad Sch Engn, Hamamatsu, Shizuoka 4328561, Japan

[2] Nagaoka Univ Technol, Nagaoka, Niigata 9402188, Japan

[3] Nanyang Technol Univ, Sch Comp Engn, Singapore, Singapore

[4] Inst Infocomm Res, Human Language Technol, Singapore, Singapore

来源：

2014 9TH INTERNATIONAL SYMPOSIUM ON CHINESE SPOKEN LANGUAGE PROCESSING (ISCSLP) | 2014年

关键词：

speech recognition; dereverberation; denoising autoencoder; environment adaptation; distant-talking speech; SPECTRAL SUBTRACTION; REVERBERATION; ADAPTATION; DOMAIN; MODEL;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

In this paper, we propose a robust distant-talking speech recognition by combining cepstral domain denoising autoencoder (DAE) and temporal structure normalization (TSN) filter. For the proposed method, after applying a DAE in the cepstral domain of speech to suppress reverberation, we apply a postprocessing technology based on temporal structure normalization (TSN) filter to reduce the noise and reverberation effects by normalizing the modulation spectra to reference spectra of clean speech. The proposed method was evaluated using speech in simulated and real reverberant environments. By combining a cepstral-domain DAE and TSN, the average Word Error Rate (WER) was reduced from 25.2% of the baseline system to 21.2% in simulated environments and from 47.5% to 41.3% in real environments, respectively.

引用

页码：379 / +

页数：3

共 31 条

[11]

Kinoshita K, 2006, INT CONF ACOUST SPEE, P817

[12]

Kinoshita K, 2013, IEEE WORK APPL SIG

[13]

Li WF, 2013, INT CONF ACOUST SPEE, P7117, DOI 10.1109/ICASSP.2013.6639043

[14]

Lincoln M, 2005, 2005 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING (ASRU), P357

[15]

Liu F.H., 1993, P 6 ARPA WORKSHOP HU, P69

[16]

Lu XG, 2013, INTERSPEECH, P436

[17]

ROBINSON T, 1995, INT CONF ACOUST SPEE, P81, DOI 10.1109/ICASSP.1995.479278

[18]

Sadjadi SO, 2011, INT CONF ACOUST SPEE, P5448

[19] Reverberation Model-Based Decoding in the Logmelspec Domain for Robust Distant-Talking Speech Recognition [J].

Sehr, Armin ;

Maas, Roland ;

Kellermann, Walter .

IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2010, 18 (07) :1676-1691

[20]

Vincent P, 2010, J MACH LEARN RES, V11, P3371

← 1 2 3 4 →