Automatic speaker recognition using a unique personal feature vector and Gaussian Mixture Models | IEEE Conference Publication | IEEE Xplore