logo Idiap Research Institute        
 [BibTeX] [Marc21]
HMM2- Extraction of Formant Features and their Use for Robust ASR
Type of publication: Idiap-RR
Citation: weber-rr-00-42
Number: Idiap-RR-42-2000
Year: 2000
Institution: IDIAP
Address: Martigny, Switzerland
Note: Published: Eurospeech 2001, Aalborg
Abstract: As recently introduced, an HMM2 can be considered as a particular case of an HMM mixture in which the HMM emission probabilities (usually estimated through Gaussian mixtures or an artificial neural network) are modeled by state-dependent, feature-based HMM (referred to as frequency HMM). A general EM training algorithm for such a structure has already been developed. Although there are numerous motivations for using such a structure, and many possible ways to exploit it, this paper will mainly focus on one particular instantiation of HMM2 in which the frequency HMM will be used to extract formant structure information, which will then be used as additional acoustic features in a standard Automatic Speech Recognition (ASR) system. While the fact that this architecture is able to automatically extract meaningful formant information is interesting by itself, empirical results will also show the robustness of these features to noise, and their potential to enhance state-of-the-art noise-robust HMM-based ASR.
Userfields: ipdinar={2000}, ipdmembership={speech}, language={English},
Projects Idiap
Authors Weber, Katrin
Bengio, Samy
Bourlard, Hervé
Crossref by weber-ar-00-42-01
Added by: [UNK]
Total mark: 0
  • rr00-42.pdf
  • rr00-42.ps.gz