ISCA Archive Interspeech 2008
ISCA Archive Interspeech 2008

On the mask modeling and feature representation in the missing-feature ASR: evaluation on the Consonant Challenge

Peter Jančovič, Munevver Kokuer

In this paper we investigate an incorporation of mask modeling within the missing-feature HMM-based ASR and also explore on feature representation within this framework. The mask model is estimated for each HMM state and mixture by using a separate Viterbi-style training procedure. We explore an employment of the frequency-filtered features and their combination with the logarithm filter-bank energies. Experimental evaluation is performed on the Consonant Challenge corpus. The obtained results show significant improvements by incorporation of the proposed methods over the standard MFT-based ASR.


doi: 10.21437/Interspeech.2008-489

Cite as: Jančovič, P., Kokuer, M. (2008) On the mask modeling and feature representation in the missing-feature ASR: evaluation on the Consonant Challenge. Proc. Interspeech 2008, 1777-1780, doi: 10.21437/Interspeech.2008-489

@inproceedings{jancovic08_interspeech,
  author={Peter Jančovič and Munevver Kokuer},
  title={{On the mask modeling and feature representation in the missing-feature ASR: evaluation on the Consonant Challenge}},
  year=2008,
  booktitle={Proc. Interspeech 2008},
  pages={1777--1780},
  doi={10.21437/Interspeech.2008-489}
}