Using prosodic and conversational features for high-performance speaker recognition: report from JHU WS'02 | IEEE Conference Publication | IEEE Xplore