16th Annual Conference of the International Speech Communication Association

Dresden, Germany
September 6-10, 2015

Non-Linear PLDA for i-Vector Speaker Verification

Sergey Novoselov, Timur Pekhovsky, Oleg Kudashev, Valentin S. Mendelev, Alexey Prudnikov

Speech Technology Center, Russia

Two approaches are presented for non-linear PLDA to be used in speaker verification. In NIST 2010 speaker recognition evaluation (SRE) tests under DET-5 conditions, the two methods and particularly their combination provided significant improvements in equal error rates and minDCF values over a standard PLDA scheme. The proposed schemes were also applied within a speaker verification system that employs DNN-based sufficient statistics calculation resulting in a 45% reduction in minDCF relative to a conventional GMM based system.

Full Paper

Bibliographic reference.  Novoselov, Sergey / Pekhovsky, Timur / Kudashev, Oleg / Mendelev, Valentin S. / Prudnikov, Alexey (2015): "Non-linear PLDA for i-vector speaker verification", In INTERSPEECH-2015, 214-218.