Optimization and evaluation of Gabor feature sets for ASR

In order to enhance automatic speech recognition performance in adverse conditions, Gabor features motivated by physiological measurements in the primary auditory cortex were optimized and evaluated. In the Aurora 2 experimental setup such localized, spectro-temporal filters combined with a Tandem system yield robust performance with a feature set size of 30. Improved results can be obtained when using a Hanning window instead of a cut-off Gaussian envelope due to better modulation frequency characteristics. An analysis of complementarity of Gabor and MFCC features shows that errors could be reduced by 55% with a perfect classifier. In a real world scenario, a relative WER reduction of 15% compared to a competitive baseline is achieved by combining the feature types, indicating the potential of this class of physiologically motivated features.