Vocal Tract Length Perturbation for Text-Dependent Speaker Verification with Autoregressive Prediction Coding

TL;DR

This paper introduces a vocal tract length perturbation technique for text-dependent speaker verification, utilizing multiple VTL factors and feature fusion, combined with autoregressive predictive coding features, to improve verification accuracy.

Contribution

The paper presents a novel VTL perturbation method combined with APC-based features for enhanced speaker verification performance.

Findings

VTL perturbation systems outperform baseline models.

Fusion of VTL-perturbed MFCC and BN features improves results.

Significant accuracy gains on the RedDots 2016 dataset.

Abstract

In this letter, we propose a vocal tract length (VTL) perturbation method for text-dependent speaker verification (TD-SV), in which a set of TD-SV systems are trained, one for each VTL factor, and score-level fusion is applied to make a final decision. Next, we explore the bottleneck (BN) feature extracted by training deep neural networks with a self-supervised objective, autoregressive predictive coding (APC), for TD-SV and compare it with the well-studied speaker-discriminant BN feature. The proposed VTL method is then applied to APC and speaker-discriminant BN features. In the end, we combine the VTL perturbation systems trained on MFCC and the two BN features in the score domain. Experiments are performed on the RedDots challenge 2016 database of TD-SV using short utterances with Gaussian mixture model-universal background model and i-vector techniques. Results show the proposed methods significantly outperform the baselines.