Presentation Attack Detection on Iris PAD Open-set (test) benchmark leaderboard · SOTA2 Research

Presentation Attack Detection on Iris PAD Open-set (test)

0.554Printout ScoreDenseNet (XENT)

Evaluation Results

Method	Links
DenseNet (XENT)2026.03		0.554	0.9377	0.8321	0.9831	0.2337	0.9167	0.9868	0.7777
DINOv2 + SVM-RBF2026.03		0.2885	0.0588	-0.518	0.0047	0.2835	-0.0118	0.0132	0.017
DINOv2 + SVM-Linear2026.03		0.1914	0.0219	-0.0924	-0.0037	0.1408	0.0035	-0.0935	0.024
DINOv2 + LogReg2026.03		0.1026	0.0255	-0.1449	-0.0187	0.0297	-0.02	-0.035	-0.0087
Equal Entropy2026.03		0.0879	-0.0535	0.0423	0.0094	0.0173	-0.0341	0.0088	0.0112
High Entropy2026.03		0.0257	0.0088	0.0166	0.0169	0.0623	-0.0294	0	0.0144
Low Entropy2026.03		-0.0403	0.0079	0.0282	0.0094	0.1676	-0.0082	0.0103	0.025
Baseline Saliency-based2026.03		-0.0522	0.0211	0.0615	0.0094	0.092	-0.0587	0.0088	0.0117
Initial Eye Tracking2026.03		-0.2701	-0.0473	0.0743	-0.043	0.1753	-0.0834	-0.3815	-0.0822
Full Eye Tracking2026.03		-0.2765	-0.021	0.0153	-0.0664	0.0824	-0.1198	-0.2353	-0.0888
De-noised Full ET2026.03		-0.2893	-0.0412	0.0717	-0.0927	0.2194	-0.1127	-0.3552	-0.0857
De-noised Initial ET2026.03		-0.3617	-0.0412	0.0551	-0.1413	0.16	-0.0599	-0.3552	-0.1063