Short Utterance Variance Modelling and Utterance Partitioning for PLDA Speaker Verification

Citations Over TimeTop 18% of 2016 papers

Abstract

This paper analyses the short utterance probabilistic linear discriminant analysis (PLDA) speaker verification with utterance partitioning and short utterance variance (SUV) modelling approaches.Experimental studies have found that instead of using single long-utterance as enrolment data, if long enrolledutterance is partitioned into multiple short utterances and average of short utterance i-vectors is used as enrolled data, that improves the Gaussian PLDA (GPLDA) speaker verification.This is because short utterance i-vectors have speaker, session and utterance variations, and utterance-partitioning approach compensates the utterance variation.Subsequently, SUV-PLDA is also studied with utterance partitioning approach, and utterancepartitioning-based SUV-GPLDA system shows relative improvement of 9% and 16% in EER for NIST 2008 and NIST 2010 truncated 10sec-10sec evaluation condition as utterancepartitioning approach compensates the utterance variation and SUV modelling approach compensates the mismatch between full-length development data and short-length evaluation data.

Related Papers

• → Primary atomic frequency standards at NIST(2001)31 cited
• Replicating UTC(NIST) at Remote Sites(2012)
• → Assessing the Impact of the National Institute of Standards and Technology’s Research Collaborations(2014)1 cited
• → Evaluation of National Institute of Standards and Technology (NIST) Thermal Conductivity Standard Reference Materials Using a 20-cm Heat Flow Meter Apparatus(1989)1 cited
• → Technique for assessing PNA measurement repeatability using a NIST standard(2005)1 cited