EnvId: A Metric Learning Approach for Forensic Few-Shot Identification of Unseen Environments

TL;DR

EnvId introduces a metric learning framework for forensic audio environment identification, enabling accurate few-shot classification of unseen environments without retraining, even with degraded or mismatched audio conditions.

Contribution

The paper presents EnvId, a novel metric learning approach for forensic audio environment identification that works in few-shot scenarios without retraining, handling challenging real-world audio conditions.

Findings

EnvId achieves high accuracy in unseen environments.

It maintains performance under audio degradations.

It outperforms baseline methods in forensic scenarios.

Abstract

Audio recordings may provide important evidence in criminal investigations. One such case is the forensic association of a recorded audio to its recording location. For example, a voice message may be the only investigative cue to narrow down the candidate sites for a crime. Up to now, several works provide supervised classification tools for closed-set recording environment identification under relatively clean recording conditions. However, in forensic investigations, the candidate locations are case-specific. Thus, supervised learning techniques are not applicable without retraining a classifier on a sufficient amount of training samples for each case and respective candidate set. In addition, a forensic tool has to deal with audio material from uncontrolled sources with variable properties and quality. In this work, we therefore attempt a major step towards practical forensic application scenarios. We propose a representation learning framework called EnvId, short for environment identification. EnvId avoids case-specific retraining by modeling the task as a few-shot classification problem. We demonstrate that EnvId can handle forensically challenging material. It provides good quality predictions even under unseen signal degradations, out-of-distribution reverberation characteristics or recording position mismatches.