Usable Speech Assignment for Speaker Identification System

Usable speech criteria are proposed to extract minimally corrupted speech for speaker identification in cochannel speech. Extracted usable segments are separated in time and need to be organized into speaker streams for speaker identification system. In this paper, we focus to organize extracted usable speech segment into a single stream for the same speaker by speaker assignment system. We extend probabilistic framework for speaker identification system to cochannel speech. We propose to use only voiced part of speech signal in training phase. In fact voiced speech segment contain the most significant information for speaker identification as opposed to other speech segment. The system is evaluated on cochannel speech and results show a significant improvement across various Target to Interferer Ratios for speaker