Open-source neural network iris matchers (TripletIris using batch-hard triplet loss and ArcIris using ArcFace loss) plus compliant C++ implementations of HDBIF and CRYPTS are released, evaluated on IREX X and eight academic datasets, and accompanied by segmentation tools to lower entry barriers for
Thirdeye: Triplet based iris recognition without normalization,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Lowering the Barrier to IREX Participation: Open-Source Algorithms, Toolkit, and Benchmarking for Iris Recognition
Open-source neural network iris matchers (TripletIris using batch-hard triplet loss and ArcIris using ArcFace loss) plus compliant C++ implementations of HDBIF and CRYPTS are released, evaluated on IREX X and eight academic datasets, and accompanied by segmentation tools to lower entry barriers for