Implementation of CORDIC based SVM for Speaker Verification System

This brief presents the implementation of a support vector machine (SVM) for speaker verification system. The proposed system comprises of a Gaussian kernel unit and a scaling unit. The proposed system can be used inside a speaker verification system for the purpose of classification of a speaker to be true or an imposer. A Gaussian kernel processing elements (GK-PEs) and an exponential processing elements are included in the support vector machine. The Gaussian kernel processing element is responsible for evaluating the kernel value of the test vector and the supporting vector. The Gaussian kernel processing element is designed to process 4 supporting vectors simultaneously. The GK-PE is designed in a pipeline fashion so as to perform 2-norm and exponential operations. The SVM makes use of an enhanced CORDIC architecture in order to calculate the exponential value. . The scaling unit is designed to perform SVM decision value evaluation. The proposed system can be used inside a speaker verification system along with a speaker feature extraction (SFE) module and a decision module. The SFE module is responsible for performing autocorrelation analysis, linear predictive coefficient (LPC) extraction, and LPC-to-cepstrum conversion. The decision module then accumulates the frame scores that are generated by all of the test frames, and is then compared with a threshold to see if the test utterance is spoken by the claimed speaker