A speaker based unsupervised speech segmentation algorithm used in conversational speech

Authors:
Yanxiang Chen;Qiong Wang
Affiliations:
College of Computer Science & Information, Hefei University of Technology, Hefei, Anhui, China and Department of Electrical Computer Engineering, University of Illinois at Urbana-Champaign, Urbana ...;College of Computer Science & Information, Hefei University of Technology, Hefei, Anhui, China
Venue:
KSEM'07 Proceedings of the 2nd international conference on Knowledge science, engineering and management
Year:
2007

Citing 3
Cited 0

DISTBIC: a speaker-based segmentation for audio data indexing

Speech Communication - Special issue on accessing information in spoken audio
Segregation of speakers for speech recognition and speaker identification

ICASSP '91 Proceedings of the Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference
A speaker tracking system based on speaker turn detection for NIST evaluation

ICASSP '00 Proceedings of the Acoustics, Speech, and Signal Processing, 2000. on IEEE International Conference - Volume 02

Quantified Score

Hi-index	0.00

Visualization

Abstract

Difference between acoustic characteristics of speakers can be applied to segment conversational speech. In this paper, an unsupervised speech segmentation algorithm is emphasized while Euclidean distance measure and the distance measure based on GLR (Generalized Likelihood Ratio) and duration model are compared. The latter measure makes use of the likelihood ratio to describe the similarity and text-independent two-speaker verification system shows it is effective in verifying segment points as the result of being sensitive to speaker changes.