Speaker diarization using data-driven audio sequencing

Author

Khemiri, Houssemeddine ; Petrovska-Delacretaz, Dijana ; Chollet, Gerard

Author_Institution

Inst. Mines-Telecom, Telecom ParisTech, Paris, France

fYear

2013

Firstpage

7736

Lastpage

7740

Abstract

In this paper, a speaker diarization system based on data-driven segmentation is proposed. In addition to the usual segmentation and clustering steps, a new module which detects repeated segments between the same shows broadcasted on different dates is added. This process is achieved by using the ALISP-based audio identification system which segments audio data into pseudo-phonetic units. The ALISP segmentation is then used to identify the similar audio segments in TV and radio shows. The system was evaluated during the ETAPE 2011 evaluation campaign and obtained a Diarization Error Rate - DER of 16.23% which was the best result among seven participants.

Keywords

audio signal processing; error statistics; speaker recognition; ALISP segmentation; ALISP-based audio identification system; DER; ETAPE 2011 evaluation; TV show; clustering steps; data-driven audio sequencing; data-driven segmentation; diarization error rate; pseudophonetic units; radio show; speaker diarization; Databases; Density estimation robust algorithm; Hidden Markov models; Mel frequency cepstral coefficient; Sequential analysis; Speech; TV; ALISP units; data-driven audio sequencing; speaker diarization;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on

Conference_Location

Vancouver, BC

ISSN

1520-6149

Type

conf

DOI

10.1109/ICASSP.2013.6639169

Filename

6639169