Acoustic models of an HMM-based classifier include various types of hidden factors such as speaker-specific characteristics and acoustic environments. If there exist a canonicalization process that represses the decrease of differences in acoustic-likelihood among categories resulted from hidden factors, a robust ASR system can be realized. We have previously proposed the canonicalization process of feature-parameters composed of three distinctive phonetic feature (DPF) extractors focused on a gender factor. This paper describes an attempt to design multiple DPF extractors corresponding to unspecific hidden factors, as well as to introduce a noise suppressor that is targeted for the canonicalization of a noise factor. In an experiment on Japanese version AURORA2 database (AURORA2-J), the proposed system achieved significant improvements when combining the canonicalization process with the noise reduction technique based on a two-stage Wiener filter.
Cite as: Fukuda, T., Ghulam, M., Nitta, T. (2005) Designing multiple distinctive phonetic feature extractors for canonicalization by using clustering technique. Proc. Interspeech 2005, 3141-3144, doi: 10.21437/Interspeech.2005-268
@inproceedings{fukuda05_interspeech, author={Takashi Fukuda and Muhammad Ghulam and Tsuneo Nitta}, title={{Designing multiple distinctive phonetic feature extractors for canonicalization by using clustering technique}}, year=2005, booktitle={Proc. Interspeech 2005}, pages={3141--3144}, doi={10.21437/Interspeech.2005-268} }