VoxBlink: A Large Scale Speaker Verification Dataset on Camera

TL;DR

VoxBlink introduces a large-scale, high-quality audio-visual speaker verification dataset with a multi-modal purification process, significantly improving speaker verification performance across various models and better reflecting real-life scenarios.

Contribution

The paper presents VoxBlink, a novel large-scale dataset with a multi-modal purification pipeline, enhancing data quality and diversity for speaker verification research.

Findings

Incorporating VoxBlink-clean improves verification accuracy by 12-30%.

VoxBlink is one of the largest publicly available speaker verification datasets.

The dataset covers real-life scenarios with data from short videos of ordinary users.

Abstract

In this paper, we introduce a large-scale and high-quality audio-visual speaker verification dataset, named VoxBlink. We propose an innovative and robust automatic audio-visual data mining pipeline to curate this dataset, which contains 1.45M utterances from 38K speakers. Due to the inherent nature of automated data collection, introducing noisy data is inevitable. Therefore, we also utilize a multi-modal purification step to generate a cleaner version of the VoxBlink, named VoxBlink-clean, comprising 18K identities and 1.02M utterances. In contrast to the VoxCeleb, the VoxBlink sources from short videos of ordinary users, and the covered scenarios can better align with real-life situations. To our best knowledge, the VoxBlink dataset is one of the largest publicly available speaker verification datasets. Leveraging the VoxCeleb and VoxBlink-clean datasets together, we employ diverse speaker verification models with multiple architectural backbones to conduct comprehensive evaluations on the VoxCeleb test sets. Experimental results indicate a substantial enhancement in performance,ranging from 12% to 30% relatively, across various backbone architectures upon incorporating the VoxBlink-clean into the training process. The details of the dataset can be found on http://voxblink.github.io