Abstract
A new class-based histogram equalization method is proposed for robust speech recognition. The proposed method aims at not only compensating for an acoustic mismatch between training and test environments but also reducing the two fundamental limitations of the conventional histogram equalization method, the discrepancy between the phonetic distributions of training and test speech data, and the nonmonotonic transformation caused by the acoustic mismatch. The algorithm employs multiple class-specific reference and test cumulative distribution functions, classifies noisy test features into their corresponding classes, and equalizes the features by using their corresponding class reference and test distributions. The minimum mean-square error log-spectral amplitude (MMSE-LSA)-based speech enhancement is added just prior to the baseline feature extraction to reduce the corruption by additive noise. The experiments on the Aurora2 database proved the effectiveness of the proposed method by reducing relative errors by over the mel-cepstral-based features and by over the conventional histogram equalization method, respectively.
Highlights
The performance of automatic speech recognition (ASR) systems degrades severely when they are employed in acoustically mismatched environments compared to the training ones
As a feature space compensation approach for robust speech recognition, the conventional histogram equalization (HEQ) technique can be effectively utilized to compensate for the acoustic mismatch between training and test environments
The conventional HEQ has two fundamental limitations caused by the mismatch of phonetic class distributions between training and test data and by the nonmonotonic transformation resulted from the acoustic mismatch
Summary
The performance of automatic speech recognition (ASR) systems degrades severely when they are employed in acoustically mismatched environments compared to the training ones. The major feature space approaches to reducing the nonlinear behaviors of the acoustic mismatch are based on the piecewise linear approximation, such as interacting multiple model (IMM) [6] and stereo-based piecewise linear compensation for environments (SPLICE) [7]. Based on the fact that HEQ is not able to compensate for the adverse effect caused by temporally random behavior of noise, we introduce the minimum mean-square error log-spectral amplitude (MMSE-LSA)-based speech enhancement technique [19] that is used as a front-end preprocessor to HEQ to further reduce the acoustic mismatch.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
