UTD-CRSS Systems for 2016 NIST Speaker Recognition Evaluation

TL;DR

This paper describes UTD-CRSS systems for the 2016 NIST SRE, focusing on domain mismatch mitigation using unlabeled in-domain data, advanced dimension reduction, and unsupervised clustering to improve speaker recognition accuracy.

Contribution

The paper introduces novel techniques for handling domain mismatch in speaker recognition by leveraging unlabeled in-domain data and advanced data processing methods.

Findings

Improved speaker recognition performance with unlabeled in-domain data.

Effective domain adaptation through data centralization and clustering.

Enhanced score calibration using unlabeled data.

Abstract

This document briefly describes the systems submitted by the Center for Robust Speech Systems (CRSS) from The University of Texas at Dallas (UTD) to the 2016 National Institute of Standards and Technology (NIST) Speaker Recognition Evaluation (SRE). We developed several UBM and DNN i-Vector based speaker recognition systems with different data sets and feature representations. Given that the emphasis of the NIST SRE 2016 is on language mismatch between training and enrollment/test data, so-called domain mismatch, in our system development we focused on: (1) using unlabeled in-domain data for centralizing data to alleviate the domain mismatch problem, (2) finding the best data set for training LDA/PLDA, (3) using newly proposed dimension reduction technique incorporating unlabeled in-domain data before PLDA training, (4) unsupervised speaker clustering of unlabeled data and using them alone or with previous SREs for PLDA training, (5) score calibration using only unlabeled data and combination of unlabeled and development (Dev) data as separate experiments.