Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION

Abstract: 

Spectrograms fusion is an effective method for incorporating complementary speech dereverberation systems. Previous linear spectrograms fusion by averaging multiple spectrograms shows outstanding performance. However, various systems with different features cannot apply this simple method. In this study, we design the minimum difference masks (MDMs) to classify the time-frequency (T-F) bins in spectrograms according to the nearest distances from labels. Then, we propose a two-stage nonlinear spectrograms fusion system for speech dereverberation. First, we conduct a multitarget learning-based speech dereverberation front-end model to get spectrograms simultaneously. Then, MDMs are estimated to take the best parts of different spectrograms. We are using spectrograms in the first stage and MDMs in the second stage to recombine T-F bins. The experiments on the REVERB challenge show that a strong feature complementarity between spectrograms and MDMs. Moreover, the proposed framework can consistently and significantly improve PESQ and SRMR, both real and simulated data, e.g., an average PESQ gain of 0.1 in all simulated data and an average SRMR gain of 1.22 in all real data.

up
0 users have voted:

Paper Details

Authors:
Hao Shi, Longbiao Wang, Meng Ge, Sheng Li, Jianwu Dang
Submitted On:
14 April 2020 - 4:41am
Short Link:
Type:
Poster
Event:
Presenter's Name:
Hao Shi
Paper Code:
ICASSP-3378
Document Year:
2020
Cite

Document Files

3378-ICASSP-2020-poster-shi

(33)

Subscribe

[1] Hao Shi, Longbiao Wang, Meng Ge, Sheng Li, Jianwu Dang, "SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION", IEEE SigPort, 2020. [Online]. Available: http://sigport.org/5093. Accessed: Jul. 13, 2020.
@article{5093-20,
url = {http://sigport.org/5093},
author = {Hao Shi; Longbiao Wang; Meng Ge; Sheng Li; Jianwu Dang },
publisher = {IEEE SigPort},
title = {SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION},
year = {2020} }
TY - EJOUR
T1 - SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION
AU - Hao Shi; Longbiao Wang; Meng Ge; Sheng Li; Jianwu Dang
PY - 2020
PB - IEEE SigPort
UR - http://sigport.org/5093
ER -
Hao Shi, Longbiao Wang, Meng Ge, Sheng Li, Jianwu Dang. (2020). SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION. IEEE SigPort. http://sigport.org/5093
Hao Shi, Longbiao Wang, Meng Ge, Sheng Li, Jianwu Dang, 2020. SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION. Available at: http://sigport.org/5093.
Hao Shi, Longbiao Wang, Meng Ge, Sheng Li, Jianwu Dang. (2020). "SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION." Web.
1. Hao Shi, Longbiao Wang, Meng Ge, Sheng Li, Jianwu Dang. SPECTROGRAMS FUSION WITH MINIMUM DIFFERENCE MASKS ESTIMATION FOR MONAURAL SPEECH DEREVERBERATION [Internet]. IEEE SigPort; 2020. Available from : http://sigport.org/5093