Dataset artefacts in anti-spoofing systems: a case study on the ASVspoof 2017 benchmark

TL;DR

This paper investigates dataset artefacts in the ASVspoof 2017 benchmark, revealing how they influence the perceived success of anti-spoofing systems and proposing methods to improve robustness and establish new baselines.

Contribution

It identifies dataset artefacts affecting anti-spoofing system performance and introduces a preprocessing approach to mitigate their impact, along with new benchmark results.

Findings

Artefacts can artificially inflate system success rates.

Discarding nonspeech segments reduces artefact exploitation.

New baseline results for frame-level and utterance-level models.

Abstract

The Automatic Speaker Verification Spoofing and Countermeasures Challenges motivate research in protecting speech biometric systems against a variety of different access attacks. The 2017 edition focused on replay spoofing attacks, and involved participants building and training systems on a provided dataset (ASVspoof 2017). More than 60 research papers have so far been published with this dataset, but none have sought to answer why countermeasures appear successful in detecting spoofing attacks. This article shows how artefacts inherent to the dataset may be contributing to the apparent success of published systems. We first inspect the ASVspoof 2017 dataset and summarize various artefacts present in the dataset. Second, we demonstrate how countermeasure models can exploit these artefacts to appear successful in this dataset. Third, for reliable and robust performance estimates on this dataset we propose discarding nonspeech segments and silence before and after the speech utterance during training and inference. We create speech start and endpoint annotations in the dataset and demonstrate how using them helps countermeasure models become less vulnerable from being manipulated using artefacts found in the dataset. Finally, we provide several new benchmark results for both frame-level and utterance-level models that can serve as new baselines on this dataset.