Individual On-Line Variance Adaptation of Frequency Filtered Parameters for Robust ASR
2006 (English)In: INTERSPEECH 2006 AND 9TH INTERNATIONAL CONFERENCE ON SPOKEN LANGUAGE PROCESSING, VOLS 1-5, 2006, 1491-1494 p.Conference paper (Refereed)
In this paper we address the problem of robust speech recognition. We propose a new method based on the individual variance adaptation of frequency filtered parameters to reduce the deleterious effects of additive narrow-band noise. The method can be interpreted as a spectral weighting that assigns increased importance to the most reliable spectral components, typically the spectral peaks. The experiments confirm that the suggested method results in significantly improved recognition rates for additive narrow-band noise.
Place, publisher, year, edition, pages
2006. 1491-1494 p.
additive noise ASR, frequency filtered parameters, spectral subtraction, model variance adaptation, spectral weighting
IdentifiersURN: urn:nbn:se:kth:diva-36102ISI: 000269965901110ScopusID: 2-s2.0-44949144177ISBN: 978-1-60423-449-7OAI: oai:DiVA.org:kth-36102DiVA: diva2:430584
9th International Conference on Spoken Language Processing/INTERSPEECH 2006 Pittsburgh, PA, 2006
QC 201107112011-07-112011-07-082011-07-11Bibliographically approved