A new classifier for speaker verification based on the fractional Brownian motion process

A novel text-independent verification system based on the fractional Brownian motion (M_dim_fBm) for automatic speaker recognition (ASR) is presented in this paper. The performance of the proposed M_dim_fBm was compared to those achieved with the GMM (Gaussian Mixture Models) classifier using the mel-cepstral coefficients. We have used a speech database - obtained from fixed and cellular phones - uttered by 75 different speakers. The results have shown the superior performance of the M_dim_fBm classifier in terms of recopition accuracy. In addition, the proposed classifier employs a much simpler modeling structure as compared to the GMM.