ISCA Archive ICSLP 2000
ISCA Archive ICSLP 2000

Improved acoustics modeling for speech recognition using transformation techniques

Carrson C. Fung, Oscar C. Au, Wanggen Wan, Chi H. Yim, Cyan L. Keung

In statistical speech recognition, misclassification often occurs when there is a mismatch between the incoming signal and the acoustics model inside the recognizer. In order to combat this problem, techniques such as Cepstral Mean Subtraction, Vocal Tract Normalization, adaptation and pronunciation model can be used.

In this paper, we proposed a new approach based on transformation technique where the output distribution function in the HMM model, a Gaussian probability density function, could be transformed to match the estimated distribution of the incoming signal by using a memoryless invertible nonlinearity function. Since the new density still has a Gaussian form, the function could be completely characterized by using the Expectation Maximization (EM) algorithm.


Cite as: Fung, C.C., Au, O.C., Wan, W., Yim, C.H., Keung, C.L. (2000) Improved acoustics modeling for speech recognition using transformation techniques. Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000), vol. 4, 176-179

@inproceedings{fung00_icslp,
  author={Carrson C. Fung and Oscar C. Au and Wanggen Wan and Chi H. Yim and Cyan L. Keung},
  title={{Improved acoustics modeling for speech recognition using transformation techniques}},
  year=2000,
  booktitle={Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000)},
  pages={vol. 4, 176-179}
}