VISER: Visually-Informed System for Enhanced Robustness in Open-Set Iris Presentation Attack Detection
Researchers have developed VISER, a new system for detecting iris presentation attacks that utilizes visually-informed priors. Experiments comparing various saliency methods, including hand annotations and eye gaze heatmaps, found that denoised eye tracking heatmaps provided the best generalization improvement. The system aims to enhance robustness in open-set iris presentation attack detection and includes trained models and code for reproducibility. AI
IMPACT Introduces a novel approach to enhance the robustness of biometric security systems against spoofing attempts.