Demographic Bias in Presentation Attack Detection of Iris Recognition Systems
This addresses bias in biometric security for users, highlighting a critical gap in PAD fairness, though it is incremental as it extends bias analysis from verification to PAD.
The paper investigates demographic bias in iris presentation attack detection (PAD) algorithms, finding that female users are significantly less protected compared to males, based on experiments using the NDCLD-2013 database.
With the widespread use of biometric systems, the demographic bias problem raises more attention. Although many studies addressed bias issues in biometric verification, there are no works that analyze the bias in presentation attack detection (PAD) decisions. Hence, we investigate and analyze the demographic bias in iris PAD algorithms in this paper. To enable a clear discussion, we adapt the notions of differential performance and differential outcome to the PAD problem. We study the bias in iris PAD using three baselines (hand-crafted, transfer-learning, and training from scratch) using the NDCLD-2013 database. The experimental results point out that female users will be significantly less protected by the PAD, in comparison to males.