Baselines and Protocols for Household Speaker Recognition

TL;DR

This paper introduces an evaluation benchmark and open-source baselines for household speaker recognition, addressing challenges like heterogeneity, short utterances, passive enrollment, and unknown speakers.

Contribution

It provides the first publicly available evaluation protocol and baseline systems for household speaker recognition tasks.

Findings

Benchmark derived from VoxCeleb and ASVspoof data

Four active enrollment algorithms included

One passive enrollment algorithm included

Abstract

Speaker recognition on household devices, such as smart speakers, features several challenges: (i) robustness across a vast number of heterogeneous domains (households), (ii) short utterances, (iii) possibly absent speaker labels of the enrollment data (passive enrollment), and (iv) presence of unknown persons (guests). While many commercial products exist, there is less published research and no publicly-available evaluation protocols or open-source baselines. Our work serves to bridge this gap by providing an accessible evaluation benchmark derived from public resources (VoxCeleb and ASVspoof 2019 data) along with a preliminary pool of open-source baselines. This includes four algorithms for active enrollment (speaker labels available) and one algorithm for passive enrollment.