A Simplified Decoding Method for a Robust Distant-Talking ASR Concept Based on Feature-Domain Dereverberation
Sehr A, Kellermann W (2008)
Publication Language: English
Publication Type: Conference contribution
Publication year: 2008
Conference Proceedings Title: Proc. Int. Workshop on Acoustic Echo and Noise Control (IWAENC)
Event location: Seattle, Washington
URI: http://www.iwaenc.org/proceedings/2008/contents/papers/9020.pdf
Abstract
A simplified decoding method for the concept of REverbera-tion MOdeling for Speech recognition (REMOS) [1] is pro-posed. In order to achieve robust distant-talking Automatic Speech Recognition (ASR), the REMOS concept uses a com-bination of clean-speech HMMs and a reverberation model to perform feature-domain dereverberation during decoding. The simplified decoding/dereverberation method proposed in this contribution significantly reduces the computational com-plexity of the concept without a major performance reduction.
Authors with CRIS profile
How to cite
APA:
Sehr, A., & Kellermann, W. (2008). A Simplified Decoding Method for a Robust Distant-Talking ASR Concept Based on Feature-Domain Dereverberation. In Proc. Int. Workshop on Acoustic Echo and Noise Control (IWAENC). Seattle, Washington, US.
MLA:
Sehr, Armin, and Walter Kellermann. "A Simplified Decoding Method for a Robust Distant-Talking ASR Concept Based on Feature-Domain Dereverberation." Proceedings of the Workshop on Acoustic Echo and Noise Control (IWAENC), Seattle, Washington 2008.
BibTeX: Download