A novel projection-based likelihood measure for noisy speech recognition

Jen-Tzung Chien*, Hsiao Chuan Wang, Lee Min Lee

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Scopus citations


The projection-based likelihood measure, an effective means of reducing noise contamination in speech recognition, dynamically searches an optimal equalization factor for adapting the cepstral mean vector of hidden Markov model (HMM) to equalize the noisy observation. In this paper, we present a novel likelihood measure which extends the adaptation mechanism to the shrinkage of covariance matrix and the adaptation bias of mean vector. A set of adaptation functions is proposed for obtaining the compensation factors. Experiments indicate that the likelihood measure proposed herein can markedly elevate the recognition accuracy.

Original languageEnglish
Pages (from-to)287-297
Number of pages11
JournalSpeech Communication
Issue number4
StatePublished - 1 Jan 1998


  • Hidden Markov model
  • Likelihood measure
  • Noise interference
  • Robustness
  • Speech recognition

Fingerprint Dive into the research topics of 'A novel projection-based likelihood measure for noisy speech recognition'. Together they form a unique fingerprint.

Cite this