Automatic speech recognition fusion approach to unsupervised speaker clustering and labeling

Author

Lawson, A.D. ; Huggins, M.C. ; Grieco, J.J. ; Galligan, S.A. ; Harris, D.M.

Author_Institution

Res. Associates for Defense Conversion, Marcy, NY

fYear

0

fDate

0-0 0

Abstract

This paper describes a fully unsupervised approach to speaker clustering and labeling employing speech recognition (ASR) technology to bootstrap speaker identification (SID). An algorithm that combined these two technologies was able to correctly cluster and label 299 NATO ship-to-ship transmissions with an accuracy of 89% in an on-line (no a priori training) scenario. This fusion approach out-performed ASR alone by 23.6%, and outperformed manually-trained VQ-SID by 12.7% and GMM/UMB-SID by 8.6%. This paper demonstrates that, under certain circumstances, unsupervised, self-organizing systems can be more effective than manually-trained ones

Keywords

marine communication; military communication; mobile radio; pattern clustering; speaker recognition; NATO ship-to-ship transmission; automatic speech recognition; speaker identification; speaker labeling; unsupervised self-organizing systems; unsupervised speaker clustering; Automatic speech recognition; Clustering algorithms; Humans; Labeling; Laboratories; Natural languages; Speaker recognition; Speech processing; Speech recognition; USA Councils;

fLanguage

English

Publisher

ieee

Conference_Titel

Aerospace Conference, 2006 IEEE

Conference_Location

Big Sky, MT

Print_ISBN

0-7803-9545-X

Type

conf

DOI

10.1109/AERO.2006.1656042

Filename

1656042