Paper Info
Title
Feature compensation employing online GMM adaptation for speech recognition in unknown severely adverse environments
Abstract
This study proposes an effective feature compensation-method to improve speech recognition in real-life speech conditions, where (i) severe background noise and channel distortion simultaneously exist, (ii) no development data is available, and (iii) clean data for ASR training and the latent clean speech in the test data are mismatched in the acoustic structure. The proposed feature compensation method employs an online GMM adaptation procedure which is based on MLLR, and a minimum statistics replacement technique for non-speech segments. The DARPA Tank corpus is used for performance evaluation, which includes severe real-life noisy conditions. The clean Broadcast News (BN) corpus is used for training the speech recognition system in this study. Experimental results show that the proposed feature compensation scheme outperforms GMM-based FMLLR and the ETSI AFE for DARPA Tank data, achieving a +5.56% relative improvement compared to FMLLR. These results demonstrate that the proposed feature compensation scheme is effective at improving speech recognition performance in unknown real-life adverse environments.
Year
DOI
Venue
2012
10.1109/ICASSP.2012.6288825
ICASSP
Keywords
Field
DocType
channel distortion,minimum statistics replacement,asr training,speech recognition,acoustic structure,online gmm adaptation,statistics replacement,regression analysis,maximum likelihood estimation,gaussian distribution,gmmadaptation,unknown severely adverse environments,latent clean speech,maximum likelihood linear regression,darpa tank corpus,real-life speech conditions,background noise,robust speech recognition,clean broadcast news corpus,etsi afe,gaussian mixture model,feature compensation,noise measurement,hidden markov models,vectors,speech
Broadcasting,Background noise,Pattern recognition,Noise measurement,Computer science,Communication channel,FMLLR,Speech recognition,Artificial intelligence,Test data,Hidden Markov model,Distortion
Conference
ISSN
ISBN
Citations 
1520-6149 E-ISBN : 978-1-4673-0044-5
978-1-4673-0044-5
3
PageRank 
References 
Authors
0.43
13
2
Name
Order
Citations
PageRank
Wooil Kim112016.95
John H. L. Hansen23215365.75