A Generalization of PLDA for Joint Modeling of Speaker Identity and Multiple Nuisance Conditions

Abstract

Probabilistic linear discriminant analysis (PLDA) is the leading method for computing scores in speaker recognition systems. The method models the vectors representing each audio sample as a sum of three terms: one that depends on the speaker identity, one that models the within-speaker variability, and one that models any remaining variability. The last two terms are assumed to be independent across samples. We recently proposed an extension of the PLDA method, which we termed Joint PLDA (JPLDA), where the second term is considered dependent on the type of nuisance condition present in the data (e.g., the language or channel). The proposed method led to significant gains for multilanguage speaker recognition when taking language as the nuisance condition. In this paper, we present a generalization of this approach that allows for multiple nuisance terms. We show results using language and several nuisance conditions describing the acoustic characteristics of the sample and demonstrate that jointly including all these factors in the model leads to better results than including only language or acoustic condition factors. Overall, we obtain relative improvements in detection cost function between 5% and 47% for various systems and test conditions with respect to standard PLDA approaches.

Short-time speaker verification with different speaking style utterances

“…In recent years, the ASV technology has achieved great success. Many session, channel compensation methods [3][4][5], new front-end [6,7], framework and speaker modeling algorithms [8][9][10][11][12] have been proposed. However, in real-world ASV applications, most technologies are still vulnerable to new domain, intra and inter-speaker variability, spoofing attacks, etc.…”