Loading...
Loading...

Go to the content (press return)

Extension of the remos concept to frequency-filtering-based features for reverberation-robust speech recognition

Author
Maas, R.; Wolf, M.; Sehr, A.; Nadeu, C.; Kellermann, W.
Type of activity
Presentation of work at congresses
Name of edition
Third Joint Workshop on Hands-free Speech Communication and Microphone Arrays (HSCMA),
Date of publication
2011
Presentation's date
2011-06
Book of congress proceedings
HSCM A ' 11: 2011 Joint Workshop on Hands-free Speech Communication and Microphone Arrays
First page
13
Last page
18
DOI
https://doi.org/10.1109/HSCMA.2011.5942381 Open in new window
Repository
http://hdl.handle.net/2117/15468 Open in new window
URL
http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=5942381&tag=1 Open in new window
Abstract
The introduction of partly decorrelated features into the REMOS (REverberationMOdeling for Speech recognition) concept for distant-talking speech recognition [1] is discussed. REMOS combines a hidden Markov model (HMM), trained on clean speech, with a reverberation model capturing certain room characteristics. The most likely contributions of both models to a reverberant observation are determined by an inner optimization problem. In HMM frameworks, decorrelated features are assumed when diagona...
Citation
Maas, R. [et al.]. Extension of the remos concept to frequency-filtering-based features for reverberation-robust speech recognition. A: Workshop on Hands-free Speech Communication and Microphone Arrays (HSCMA),. "2011 Joint Workshop on Hands-free Speech Communication and Microphone Arrays". Edinburgh: 2011, p. 13-18.
Group of research
IDEAI-UPC - Intelligent Data Science and Artificial Intelligence Research Center
TALP - Centre for Language and Speech Technologies and Applications
VEU - Speech Processing Group

Participants

  • Maas, Roland  (author and speaker )
  • Wolf, Martin  (author and speaker )
  • Sehr, Armin  (author and speaker )
  • Nadeu Camprubí, Climent  (author and speaker )
  • Kellermann, Walter  (author and speaker )