Phonetic speaker recognition using maximum-likelihood binary-decision tree models | IEEE Conference Publication | IEEE Xplore