A noise robust front-end for speech recognition using hough transform and cumulative distribution mapping

This paper describes a novel and noise robust front-end that employs the use of Hough transform for simultaneous frequency and temporal masking, together with cumulative distribution mapping of cepstral coefficients, for noisy speech recognition. Recognition experiments on the Aurora II connected digits database have revealed that the proposed front-end achieves an average digit recognition accuracy of 83.67%. Compared with the recognition results obtained by using the ETSI standard Mel-cepstral front-end, this accuracy represents a relative error rate reduction of around 58%.