Employment of Subspace Gaussian Mixture Models in speaker recognition | IEEE Conference Publication | IEEE Xplore