Exploiting models intrinsic robustness for noisy speech recognition

Christophe Cerisara 1 Dominique Fohr 1 Odile Mella 1 Irina Illina 1
1 PAROLE - Analysis, perception and recognition of speech
INRIA Lorraine, LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
Abstract : We propose in this paper an original approach to build masks in the framework of missing data recognition. The proposed soft masks are estimated from the models themselves, and not from the test signal as it is usually the case. They represent the intrinsic robustness of model's log-spectral coefficients. The method is validated with cepstral models, on two synthetic and two real-life noises, at different signal-to-noise ratios. We further discuss how such masks can be combined with other signal-based masks and noise compensation techniques.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [10 references]  Display  Hide  Download

https://hal.inria.fr/inria-00099890
Contributor : Publications Loria <>
Submitted on : Tuesday, September 26, 2006 - 10:03:33 AM
Last modification on : Friday, February 9, 2018 - 1:20:05 PM
Document(s) archivé(s) le : Wednesday, March 29, 2017 - 12:53:10 PM

Identifiers

  • HAL Id : inria-00099890, version 1

Collections

Citation

Christophe Cerisara, Dominique Fohr, Odile Mella, Irina Illina. Exploiting models intrinsic robustness for noisy speech recognition. 8th International Conference on Spoken Language Processing - ICSLP'2004, 2004, Jeju, Corée du Sud, 4 p. ⟨inria-00099890⟩

Share

Metrics

Record views

197

Files downloads

32