Speaker Diarization using Normalized Cross Likelihood Ratio - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2007

Speaker Diarization using Normalized Cross Likelihood Ratio

Résumé

In this paper, we present the Normalized Cross Likelihood Ratio (NCLR) and the advantages of using it in a speaker diarization system. First, the NCLR is used as a dissimilarity measure between two Gaussian speaker models in the speaker change detection step and its contribution to the performance of speaker change detection is compared with those of BIC and Hostelling's T2-Statistic measures. Then, the NCLR measure is modified to deal with multi-gaussian adapted models in the cluster recombination step. This step ends the step-by-step speaker diarization process after the BIC-based hierarchical clustering and the Viterbi re-segmentation steps. By comparing the NCLR measure with the CLR (Cross Likelihood Ratio) one, more than 30% of relative diarization error is reduced in ESTER evaluation data.
Fichier non déposé

Dates et versions

hal-00163855 , version 1 (18-07-2007)

Identifiants

  • HAL Id : hal-00163855 , version 1

Citer

Viet-Bac Le, Odile Mella, Dominique Fohr. Speaker Diarization using Normalized Cross Likelihood Ratio. INTERSPEECH 2007, Aug 2007, Antwerp, Belgium. pp.4. ⟨hal-00163855⟩
255 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More