On compensating the Mel-frequency cepstral coefficients for noisy speech recognition

This paper describes a novel noise-robust automatic speech recognition (ASR) front-end that employs a combination of Mel-filterbank output compensation and cumulative distribution mapping of cepstral coefficients with truncated Gaussian distribution. Recognition experiments on the Aurora II connected digits database reveal that the proposed front-end achieves an average digit recognition accuracy of 84.92% for a model set trained from clean speech data. Compared with the ETSI standard Mel-cepstral front-end, the proposed front-end is found to obtain a relative error rate reduction of around 61%. Moreover, the proposed front-end can provide comparable recognition accuracy with the ETSI advanced front-end, at less than half the computation load.