Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Eye Sclera for Fair Face Image Quality Assessment

Wassim Kabbani, Kiran Raja, Raghavendra Ramachandra, Christoph Busch

TL;DR

The paper addresses fairness in Face Image Quality Assessment (FIQA) across skin tones, a challenge when quality metrics rely on skin-based cues. It investigates the eye sclera as a demographically neutral region for assessing quality components, adapting ISO/IEC 29794-5–style measures for dynamic range and exposure to sclera input and evaluating them on FRLL-derived data with synthetic variations, using ArcFace and Error-vs-Discard Characteristic (EDC) analyses. Findings show sclera-based quality assessments achieve comparable predictive power to skin-based measures for dynamic range and exposure components, with EDC curves indicating skin-tone agnostic performance. The work suggests sclera-based FIQA as a practical, bias-reducing supplement or alternative to traditional skin-based metrics, aligning with standardization efforts and enabling fairer face recognition deployments.

Abstract

Fair operational systems are crucial in gaining and maintaining society's trust in face recognition systems (FRS). FRS start with capturing an image and assessing its quality before using it further for enrollment or verification. Fair Face Image Quality Assessment (FIQA) schemes therefore become equally important in the context of fair FRS. This work examines the sclera as a quality assessment region for obtaining a fair FIQA. The sclera region is agnostic to demographic variations and skin colour for assessing the quality of a face image. We analyze three skin tone related ISO/IEC face image quality assessment measures and assess the sclera region as an alternative area for assessing FIQ. Our analysis of the face dataset of individuals from different demographic groups representing different skin tones indicates sclera as an alternative to measure dynamic range, over- and under-exposure of face using sclera region alone. The sclera region being agnostic to skin tone, i.e., demographic factors, provides equal utility as a fair FIQA as shown by our Error-vs-Discard Characteristic (EDC) curve analysis.

Eye Sclera for Fair Face Image Quality Assessment

TL;DR

The paper addresses fairness in Face Image Quality Assessment (FIQA) across skin tones, a challenge when quality metrics rely on skin-based cues. It investigates the eye sclera as a demographically neutral region for assessing quality components, adapting ISO/IEC 29794-5–style measures for dynamic range and exposure to sclera input and evaluating them on FRLL-derived data with synthetic variations, using ArcFace and Error-vs-Discard Characteristic (EDC) analyses. Findings show sclera-based quality assessments achieve comparable predictive power to skin-based measures for dynamic range and exposure components, with EDC curves indicating skin-tone agnostic performance. The work suggests sclera-based FIQA as a practical, bias-reducing supplement or alternative to traditional skin-based metrics, aligning with standardization efforts and enabling fairer face recognition deployments.

Abstract

Fair operational systems are crucial in gaining and maintaining society's trust in face recognition systems (FRS). FRS start with capturing an image and assessing its quality before using it further for enrollment or verification. Fair Face Image Quality Assessment (FIQA) schemes therefore become equally important in the context of fair FRS. This work examines the sclera as a quality assessment region for obtaining a fair FIQA. The sclera region is agnostic to demographic variations and skin colour for assessing the quality of a face image. We analyze three skin tone related ISO/IEC face image quality assessment measures and assess the sclera region as an alternative area for assessing FIQ. Our analysis of the face dataset of individuals from different demographic groups representing different skin tones indicates sclera as an alternative to measure dynamic range, over- and under-exposure of face using sclera region alone. The sclera region being agnostic to skin tone, i.e., demographic factors, provides equal utility as a fair FIQA as shown by our Error-vs-Discard Characteristic (EDC) curve analysis.
Paper Structure (6 sections, 9 figures, 1 table, 3 algorithms)

This paper contains 6 sections, 9 figures, 1 table, 3 algorithms.

Table of Contents

  1. Introduction
  2. Related Works
  3. Methods
  4. Dynamic Range
  5. Under and Over Exposure
  6. Conclusion and Future Work

Figures (9)

  • Figure 1: Images from FRLL dataset b7
  • Figure 2: Examples of non-compliant images due to low dynamic range
  • Figure 3: Dynamic range EDC curves for face vs. sclera with a starting error rate of 0.05 and ArcFace as face recognition system.
  • Figure 4: Normal vs. Low dynamic range for face skin vs. sclera
  • Figure 5: Examples of non-compliant exposure
  • ...and 4 more figures