Adaptation of PLDA for multi-source text-independent speaker verification

Abstract

Probabilistic linear discriminant analysis (PLDA) is widely described as an effective model for text-independent speaker verification in the i-vector space. The PLDA scoring function is typically formulated as the likelihood ratio between the speaker-adapted and the universal PLDAs. In this case, the adaptation of PLDA was performed through the speaker… (More)
DOI: 10.1109/ICASSP.2017.7953184

Topics

2 Figures and Tables

Slides referencing similar topics