This paper presents a simplified post-classification framework for enhancing the performance of a given speaker recognition classifier by means of other "auxiliary" classifiers. We call it Virtual Fusion, since the assisting classifiers are used only for training the post-classifier and are not necessary in operating mode. Experiments performed using Nist'04 and '05 evaluations suggest that the proposed technique is able to consistently improve the EER of a typical GMM-cepstrum classifier by up to 15%.
Bibliographic reference. Solewicz, Yosef A. / Koppel, Moshe (2007): "Virtual fusion for speaker recognition", In INTERSPEECH-2007, 1997-2000.