DMCA
AN INVESTIGATION ON BACK-END FOR SPEAKER RECOGNITION IN MULTI-SESSION ENROLLMENT
| Citations: | 10 - 6 self |
BibTeX
@MISC{Liu_aninvestigation,
author = {Gang Liu and Taufiq Hasan and Hynek Bořil and John H. L. Hansen},
title = {AN INVESTIGATION ON BACK-END FOR SPEAKER RECOGNITION IN MULTI-SESSION ENROLLMENT},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
This study explores various back-end classifiers for robust speaker recognition in multi-session enrollment, with emphasis on optimal utilization and organization of speaker information present in the development data. Our objective is to construct a highly discriminative back-end framework by fusing several back-ends on an i-vector system framework. It is demonstrated that, by using different information/data configuration and modeling schemes, performance of the fused system can be significantly improved compared to an individual system using a single front-end and back-end. Averaged across both genders, we obtain a relative improvement in EER and minDCF by 56.5 % and 49.4%, respectively. Consistent performance gains obtained using the proposed strategy validates its effectiveness. This system is part of the CRSS ’ NIST SRE 2012 submission system.
Keyphrases
speaker information present cr nist sre optimal utilization submission system discriminative back-end framework robust speaker recognition different information data configuration modeling scheme individual system multi-session enrollment various back-end classifier several back-ends development data relative improvement fused system consistent performance gain single front-end i-vector system framework
