Output details
15 - General Engineering
University of Cambridge
Return to search
Output 0 of 0 in the submission
Article title
Noisy constrained maximum-likelihood linear regression for noise-robust speech recognition
Type
D - Journal article
Title of journal
IEEE Transactions on Audio, Speech and Language Processing
Article number
-
Volume number
19
Issue number
2
First page of article
315
ISSN of journal
1558-7916
Year of publication
2011
Number of additional authors
1
Additional information
The scheme described in this paper (and other noise compensation schemes) led to an invitation to join the BBN-team bidding for funding under the DARPA-funded Robust Automatic Transcription of Speech (RATS) program, and the award of a research grant with the BBN-led team. This project aims to improve performance of speech systems under degraded communication channels, using an extension of the standard approach to adapting a speech recognition system to a speaker used in state-of-the-art speech recognition systems (for example the research systems produced by IBM, SRI, BBN), as developed in this paper.
Interdisciplinary
-
Cross-referral requested
-
Research group
None
Proposed double-weighted
No
Double-weighted statement
-
Reserve for a double-weighted output
No
Non-English
No
English abstract
-