Impact of Iris Pigmentation on Performance Bias in Visible Iris Verification Systems: A Comparative Study

TL;DR

The results indicate that iris recognition systems generally exhibit higher accuracy for blue irises compared to dark irises, and underscore the need for more inclusive dataset collection and model refinement to reduce bias and promote equitable biometric recognition across varying iris pigmentation and device configurations.

Abstract

Iris recognition technology plays a critical role in biometric identification systems, but their performance can be affected by variations in iris pigmentation. In this work, we investigate the impact of iris pigmentation on the efficacy of biometric recognition systems, focusing on a comparative analysis of blue and dark irises. Data sets were collected using multiple devices, including P1, P2, and P3 smartphones [4], to assess the robustness of the systems in different capture environments [19]. Both traditional machine learning techniques and deep learning models were used, namely Open-Iris, ViT-b, and ResNet50, to evaluate performance metrics such as Equal Error Rate (EER) and True Match Rate (TMR). Our results indicate that iris recognition systems generally exhibit higher accuracy for blue irises compared to dark irises. Furthermore, we examined the generalization capabilities of these systems across different iris colors and devices, finding that while training on diverse datasets enhances recognition performance, the degree of improvement is contingent on the specific model and device used. Our analysis also identifies inherent biases in recognition performance related to iris color and cross-device variability. These findings underscore the need for more inclusive dataset collection and model refinement to reduce bias and promote equitable biometric recognition across varying iris pigmentation and device configurations.

Figures (5)

• Figure 1: Illustrates an iris recognition system comparing blue and dark irises, using similarity scores and analyzing fairness with DPD and EOD metrics to address demographic biases.
• Figure 2: Illustration of the iris recognition system's performance on a dataset with varying iris pigmentation (dark and blue irises) using three different models: Open-Iris, ViT-b, and ResNet-50. The study evaluates how iris pigmentation influences recognition accuracy and assesses fairness through Demographic Parity Difference (DPD) and Equalized Odds Difference (EoD) metrics to identify potential system biases and demographic disparities.
• Figure 3: Preprocessed blue and dark irises showing reduced pigmentation differences and noise removal (white pixels), highlighting uniformity for biometric analysis.
• Figure 4: DET curve for comparative analysis among three different models. The X-axis indicates the false match rate and the y-axis indicates the false non-match rate of the Iris Pigmentation (Blue color) dataset.
• Figure 5: DET curve for comparative analysis among three different models. The X-axis indicates the false match rate and the y-axis indicates the false non-match rate of the Iris Pigmentation (Dark color) dataset.