Are Face Detection Models Biased?

TL;DR

This paper investigates bias in face detection models by analyzing facial localization performance across demographic groups using a newly curated dataset, revealing significant disparities and confounding factors.

Contribution

Introduces the F2LA dataset with detailed annotations for bias analysis in face detection, focusing on facial localization rather than binary classification.

Findings

Detection accuracy varies significantly across gender and skin-tone.

Confounding factors influence detection disparities beyond demographic attributes.

F2LA dataset enables comprehensive bias analysis in face detection models.

Abstract

The presence of bias in deep models leads to unfair outcomes for certain demographic subgroups. Research in bias focuses primarily on facial recognition and attribute prediction with scarce emphasis on face detection. Existing studies consider face detection as binary classification into 'face' and 'non-face' classes. In this work, we investigate possible bias in the domain of face detection through facial region localization which is currently unexplored. Since facial region localization is an essential task for all face recognition pipelines, it is imperative to analyze the presence of such bias in popular deep models. Most existing face detection datasets lack suitable annotation for such analysis. Therefore, we web-curate the Fair Face Localization with Attributes (F2LA) dataset and manually annotate more than 10 attributes per face, including facial localization information. Utilizing the extensive annotations from F2LA, an experimental setup is designed to study the performance of four pre-trained face detectors. We observe (i) a high disparity in detection accuracies across gender and skin-tone, and (ii) interplay of confounding factors beyond demography. The F2LA data and associated annotations can be accessed at http://iab-rubric.org/index.php/F2LA.