Deep Neural Network Based Monaural Speech Enhancement with Low-Rank Analysis and Speech Present Probability
Summary :
A monaural speech enhancement method combining deep neural network (DNN) with low rank analysis and speech present probability is proposed in this letter. Low rank and sparse analysis is first applied on the noisy speech spectrogram to get the approximate low rank representation of noise. Then a joint feature training strategy for DNN based speech enhancement is presented, which helps the DNN better predict the target speech. To reduce the residual noise in highly overlapping regions and high frequency domain, speech present probability (SPP) weighted post-processing is employed to further improve the quality of the speech enhanced by trained DNN model. Compared with the supervised non-negative matrix factorization (NMF) and the conventional DNN method, the proposed method obtains improved speech enhancement performance under stationary and non-stationary conditions.
- Publication
- IEICE TRANSACTIONS on Fundamentals Vol.E101-A No.3 pp.585-589
- Publication Date
- 2018/03/01
- Publicized
- Online ISSN
- 1745-1337
- DOI
- 10.1587/transfun.E101.A.585
- Type of Manuscript
- LETTER
- Category
- Noise and Vibration
Authors
Wenhua SHI
the Army Engineering University of PLA
Xiongwei ZHANG
the Army Engineering University of PLA
Xia ZOU
the Army Engineering University of PLA
Meng SUN
the Army Engineering University of PLA
Wei HAN
the Army Engineering University of PLA
Li LI
the Army Engineering University of PLA
Gang MIN
National University of Defense Technology
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Wenhua SHI, Xiongwei ZHANG, Xia ZOU, Meng SUN, Wei HAN, Li LI, Gang MIN, "Deep Neural Network Based Monaural Speech Enhancement with Low-Rank Analysis and Speech Present Probability" in IEICE TRANSACTIONS on Fundamentals,
vol. E101-A, no. 3, pp. 585-589, March 2018, doi: 10.1587/transfun.E101.A.585.
Abstract: A monaural speech enhancement method combining deep neural network (DNN) with low rank analysis and speech present probability is proposed in this letter. Low rank and sparse analysis is first applied on the noisy speech spectrogram to get the approximate low rank representation of noise. Then a joint feature training strategy for DNN based speech enhancement is presented, which helps the DNN better predict the target speech. To reduce the residual noise in highly overlapping regions and high frequency domain, speech present probability (SPP) weighted post-processing is employed to further improve the quality of the speech enhanced by trained DNN model. Compared with the supervised non-negative matrix factorization (NMF) and the conventional DNN method, the proposed method obtains improved speech enhancement performance under stationary and non-stationary conditions.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.E101.A.585/_p
Copy
@ARTICLE{e101-a_3_585,
author={Wenhua SHI, Xiongwei ZHANG, Xia ZOU, Meng SUN, Wei HAN, Li LI, Gang MIN, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Deep Neural Network Based Monaural Speech Enhancement with Low-Rank Analysis and Speech Present Probability},
year={2018},
volume={E101-A},
number={3},
pages={585-589},
abstract={A monaural speech enhancement method combining deep neural network (DNN) with low rank analysis and speech present probability is proposed in this letter. Low rank and sparse analysis is first applied on the noisy speech spectrogram to get the approximate low rank representation of noise. Then a joint feature training strategy for DNN based speech enhancement is presented, which helps the DNN better predict the target speech. To reduce the residual noise in highly overlapping regions and high frequency domain, speech present probability (SPP) weighted post-processing is employed to further improve the quality of the speech enhanced by trained DNN model. Compared with the supervised non-negative matrix factorization (NMF) and the conventional DNN method, the proposed method obtains improved speech enhancement performance under stationary and non-stationary conditions.},
keywords={},
doi={10.1587/transfun.E101.A.585},
ISSN={1745-1337},
month={March},}
Copy
TY - JOUR
TI - Deep Neural Network Based Monaural Speech Enhancement with Low-Rank Analysis and Speech Present Probability
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 585
EP - 589
AU - Wenhua SHI
AU - Xiongwei ZHANG
AU - Xia ZOU
AU - Meng SUN
AU - Wei HAN
AU - Li LI
AU - Gang MIN
PY - 2018
DO - 10.1587/transfun.E101.A.585
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E101-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2018
AB - A monaural speech enhancement method combining deep neural network (DNN) with low rank analysis and speech present probability is proposed in this letter. Low rank and sparse analysis is first applied on the noisy speech spectrogram to get the approximate low rank representation of noise. Then a joint feature training strategy for DNN based speech enhancement is presented, which helps the DNN better predict the target speech. To reduce the residual noise in highly overlapping regions and high frequency domain, speech present probability (SPP) weighted post-processing is employed to further improve the quality of the speech enhanced by trained DNN model. Compared with the supervised non-negative matrix factorization (NMF) and the conventional DNN method, the proposed method obtains improved speech enhancement performance under stationary and non-stationary conditions.
ER -
English
