The IRISA/ELISA Speaker Detection and Tracking Systems for the NIST'99 Evaluation Campaign
| Autor: | Mouhamadou Seck, Frédéric Bimbot, Raphaël Blouet |
|---|---|
| Rok vydání: | 2000 |
| Předmět: |
business.industry
Computer science Applied Mathematics Speech recognition Tracking system Speaker recognition Mixture model Speaker diarisation Computational Theory and Mathematics Artificial Intelligence Signal Processing Maximum a posteriori estimation NIST Digital signal Computer Vision and Pattern Recognition Electrical and Electronic Engineering Statistics Probability and Uncertainty business Change detection |
| Zdroj: | Digital Signal Processing. 10:154-171 |
| ISSN: | 1051-2004 |
| DOI: | 10.1006/dspr.1999.0364 |
| Popis: | Seck, Mouhamadou, Blouet, Raphael, and Bimbot, Frederic, The IRISA/ELISA Speaker Detection and Tracking Systems for the NIST'99 Evaluation Campaign, Digital Signal Processing10(2000), 154?171.This article presents the various systems developed by IRISA, around the ELISA platform, for the NIST'99 evaluation campaign in speaker detection and tracking. The main features of these systems are the implementation of a Maximum A Posteriori approach for speaker model estimation, an utterance-length dependent z-normalization scheme for test segment scoring, a speaker/world mixture model for addressing the two-speaker detection task, and the use of a change point detection method for speaker tracking. The performance of the various systems on the NIST'99 evaluation data are reported. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect