Using post-classifiers to enhance fusion of low- and high-level speaker recognition

Yosef A. Solewicz, Moshe Koppel

Research output: Contribution to journalArticlepeer-review

7 Scopus citations


This paper proposes a method for automatic correction of bias in speaker recognition systems, especially fusion-based systems. The method is based on a post-classifier which learns the relative performance obtained by the constituent systems in key trials, given the training and testing conditions in which they occurred. These conditions generally reflect train/test mismatch in factors such as channel, noise, speaker stress, etc. Results obtained with several state-of-the-art systems showed up to 20% decrease in EER compared to ordinary fusion in the NIST'05 Speaker Recognition Evaluation.

Original languageEnglish
Article number4291613
Pages (from-to)2063-2071
Number of pages9
JournalIEEE Transactions on Audio, Speech and Language Processing
Issue number7
StatePublished - Sep 2007


  • Fusion
  • Machine learning
  • Post-classification
  • Speaker recognition


Dive into the research topics of 'Using post-classifiers to enhance fusion of low- and high-level speaker recognition'. Together they form a unique fingerprint.

Cite this