(Un)fair Exposure in Deep Face Rankings at a Distance

TL;DR

This paper investigates demographic biases in deep face ranking systems used in law enforcement, revealing significant exposure disparities across demographic groups and emphasizing the need for corrective policies.

Contribution

It introduces a novel experimental framework with multiple face encoders and datasets to analyze biases in forensic face rankings, an underexplored area.

Findings

Biases in exposure are prevalent across tested face encoders.

Current face ranking methods do not adequately counteract demographic biases.

Biases affect both re-identification and identification tasks.

Abstract

Law enforcement regularly faces the challenge of ranking suspects from their facial images. Deep face models aid this process but frequently introduce biases that disproportionately affect certain demographic segments. While bias investigation is common in domains like job candidate ranking, the field of forensic face rankings remains underexplored. In this paper, we propose a novel experimental framework, encompassing six state-of-the-art face encoders and two public data sets, designed to scrutinize the extent to which demographic groups suffer from biases in exposure in the context of forensic face rankings. Through comprehensive experiments that cover both re-identification and identification tasks, we show that exposure biases within this domain are far from being countered, demanding attention towards establishing ad-hoc policies and corrective measures. The source code is available at https://github.com/atzoriandrea/ijcb2023-unfair-face-rankings