Parametric Cepstral Mean Normalization for Robust Speech Recognition

This paper proposes a new channel normalization algorithm called parametric cepstral mean normalization (PCMN) to increase robustness of speech recognition to varying acoustic conditions. Rather than using a simple average of input speech features as channel estimate, as done in the traditional CMN,...

Full description

Saved in:
Bibliographic Details
Published inICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) pp. 6735 - 6739
Main Authors Kalinli, Ozlem, Bhattacharya, Gautam, Weng, Chao
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.05.2019
Subjects
Adaptation models
Cepstral analysis
cepstral mean normalization
channel normalization
Frequency dependence
Indexes
Noise measurement
Robust automatic speech recognition
Training
Online AccessGet full text

Cover

More Information
Summary:This paper proposes a new channel normalization algorithm called parametric cepstral mean normalization (PCMN) to increase robustness of speech recognition to varying acoustic conditions. Rather than using a simple average of input speech features as channel estimate, as done in the traditional CMN, PCMN weighs the running average of input speech frames in a frequency dependent manner. These weights are jointly optimized together with parameters of the acoustic model training. Experimental results show that, in contrast to traditional CMN, which degrades performance on clean data, PCMN provides 5% relative improvement on clean data, while also providing 11.2% relative improvement on far-field test data. We also propose an adaptive version of PCMN, called aPCMN, where both input speech features and channel estimates have weights. These weights are computed at run time and they change dynamically based on the input speech. aPCMN provides 13.0% relative improvement on far-field test set, while still maintaining 5% relative improvement on clean data.
ISSN:2379-190X
DOI:10.1109/ICASSP.2019.8683674