FunFace: Feature Utility and Norm Estimation for Face Recognition

TL;DR

FunFace introduces a novel adaptive margin loss for face recognition that incorporates biometric utility estimates, improving performance especially on low-quality face samples.

Contribution

The paper proposes FunFace, a new loss function that integrates biometric utility into face recognition training, enhancing robustness across varying sample qualities.

Findings

FunFace achieves competitive results on high-quality benchmarks.

FunFace surpasses state-of-the-art models on low-quality face benchmarks.

Incorporating utility estimates improves face recognition robustness.

Abstract

Face Recognition (FR) is used in a variety of application domains, from entertainment and banking to security and surveillance. Such applications rely on the FR model to be robust and perform well in a variety of settings. To achieve this, state-of-the-art FR models typically use expressive adaptive margin loss functions, which tie the feature norm to concepts related to sample quality, such as recognizability and perceptual image quality. Recently, through the development of Face Image Quality Assessment (FIQA) techniques, biometric utility has become the preferred measure of face-image quality and has been shown to be a better predictor of the usefulness of samples for face recognition compared to more human-centric aspects, such as resolution, blur, and lighting, tied to general image quality. While image quality expressed through feature norms exhibits a certain level of correlation with biometric utility, it does not fully encapsulate all aspects of utility. To address this point, we propose a new adaptive margin loss, FunFace (Face Recognition Through Utility and Norm Estimation), which incorporates biometric utility, estimated by the Certainty Ratio, into the adaptive margin, taking inspiration from AdaFace. We show that FunFace (when used to train a face recognition model) achieves competitive results to other state-of-the-art FR models on benchmarks containing high-quality samples, while surpassing them on low quality benchmarks.