Segmentation of speech using speaker identification

40 Years of Progress in Automatic Speaker Recognition

“…(6) Relation to other speech research: Speaker characterization techniques were related to research on improving speech recognition accuracy by speaker adaptation [16], improving synthesized speech quality by adding the natural characteristics of voice individuality, and converting synthesized voice individuality from one speaker to another. Studies on speaker diarization, that is, automatically extracting the speech periods of each person separately ("who spoke when") from a dialogue/conversation/meeting involving more than two people appeared as an extension of speaker recognition technology [21,45,49]. Speaker segmentation and clustering techniques have been used to aid in the adaptation of speech recognizers and for supplying metadata for audio indexing and searching.…”

Automatic Summarization

“…Diarization is the task of automatically identifying sections of spoken audio and correctly labeling them with their characteristics, for example, speech, non-speech, male-speech, female-speech, music, noise. Although speaker identification played a role in early segmentation approaches, e.g., [300], determination of the identity of the speaker, called speaker identification, or confirmation of a presumed speaker identity, called speaker verification, does not fall into the scope of the diarization task.…”

Approaches to Speaker Detection and Tracking in Conversational Speech

“…In this paper we use a blind clustering approach described in [12] to generate homogeneous regions with no prior knowledge of the hypothesized speaker. For speaker detection, we score each homogeneous region as in the single-speaker case and then take the maximum score as the overall detection score.…”