Modulation Spectrum Equalization for Improved Robust Speech Recognition

2013 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference

2013

In this paper, we present a novel approach to enhancing the speech features in the modulation spectrum for better recognition performance in noise-corrupted environments. In the presented approach, termed modulation spectrum powerlaw expansion (MSPLE), the speech feature temporal stream is first pre-processed by some statistics compensation technique, such as mean and variance normalization (MVN), cepstral gain normalization (CGN) and MVN plus ARMA filtering (MVA), and then the magnitude part of the modulation spectrum (Fourier transform) for the feature stream is raised to a power (exponentiated). We find that MSPLE can highlight the speech components and reduce the noise distortion existing in the statistics-compensated speech features. With the Aurora-2 digit database task, experimental results reveal that the above process can consistently achieve very promising recognition accuracy under a wide range of noise-corrupted environments. MSPLE operated on MVN-preprocessed features brings about 55% in error rate reduction relative to the MFCC baseline and significantly outperforms the single MVN. Furthermore, performing MSPLE on the lower sub-band modulation spectra gives the results very close to those from the full-band modulation spectra updated by MSPLE, indicating that a lesscomplicated MSPLE suffices to produce noise-robust speech features.

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Modulation spectrum power-law expansion for robust speech recognition

Fan

2013 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference

2013

“…To enhance the recognition performance robustness, a lot of approaches have been proposed [2] [3] [4]. A successful category of approaches aims to produce robust features that are less sensitive to environmental mismatch between training and testing conditions.…”

Section: Introductionmentioning

confidence: 99%

Filtering on the temporal probability sequence in histogram equalization for robust speech recognition

Wang

2013 IEEE International Conference on Acoustics, Speech and Signal Processing

2013

In this paper, we propose a filter-based histogram equalization (FHEQ) approach for robust speech recognition. The FHEQ approach first represents the original acoustic feature sequence with statistic probability. Then, a temporal average (TA) filter is applied to smooth the statistic probability sequence. Finally, the filtered statistic probability sequence is transformed to form a new acoustic feature stream. Filtering on statistic probability of a feature sequence is a novel concept that can incorporate the advantages of the conventional histogram equalization (HEQ) and temporal filtering techniques for better noise robustness. Our experimental results on the Aurora-2 and Aurora-4 tasks show that FHEQ outperforms the conventional cepstral mean subtraction (CMS), cepstral mean and variance normalization (CMVN), and HEQ. Furthermore, we conducted a comparison test on TA-HEQ and HEQ-TA, which apply a TA filter to smooth acoustic features before and after the HEQ processing, respectively. The test results show that FHEQ outperforms both TA-HEQ and HEQ-TA, suggesting that filtering in probability is more effective than filtering in acoustic feature.

“…The idea of moment normalization in CMS, CMVN and CHN stated above has been adopted to compensate the modulation spectrum [4][5][6][7], which is the discrete Fourier transform of the cepstral time series of an utterance. For instance, in the method of spectral histogram equalization (SHE) [4], all the (magnitude) spectral points in the entire modulation frequency band are treated in a joint manner to estimate a histogram, which is then used to produce a transform for the modulation spectrum and to make the updated modulation spectrum close to a target histogram.…”

Section: Introductionmentioning

confidence: 99%

“…For instance, in the method of spectral histogram equalization (SHE) [4], all the (magnitude) spectral points in the entire modulation frequency band are treated in a joint manner to estimate a histogram, which is then used to produce a transform for the modulation spectrum and to make the updated modulation spectrum close to a target histogram. SHE can be viewed as the spectral-domain counterpart of CHN, and it effectively reduces the noise effect in the modulation spectrum of cepstral time series.…”

Section: Introductionmentioning

confidence: 99%

Overlapped sub-band modulation spectrum normalization techniques for robust speech recognition

Fan

Yeh

2013 10th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD)

2013

This paper proposes a novel approach to enhance the speech features in noise robustness for speech recognition. In the proposed approach, the speech feature time sequence is first converted into the modulation spectral domain via discrete Fourier transform (DFT). The magnitude part of the modulation spectrum is decomposed into overlapped non-uniform sub-band segments, and then each sub-band segment is individually processed by a specific statistics normalization method, like mean and variance normalization (MVN) and histogram equalization (HEQ). Finally, we reconstruct the feature time sequence with all the modified sub-band magnitude spectral segments together with the original phase spectrum using the inverse DFT. During the process, the components that correspond to more important modulation spectral bands in the feature sequence can be processed separately, and more spectral samples within each band give rise to more accurate statistic estimates due to overlapping the adjacent segments. For the Aurora-2 clean-condition training task, the new proposed method gives rise to significant improvement in recognition accuracy over the baseline results, and it behaves better than the similar technique dealing with non-overlapped sub-bands.