Accurate Detection of Non-Iris Occlusions

Accurate detection of iris eyelids and reflections is the prerequisite for the accurate iris recognition, both in near-infrared or visible spectrum measurements. Undected iris occlusions otherwise dramatically decrease the iris recognition rate. This paper presents a fast multispectral iris occlusions detection method based on the underlying multispectral spatial probabilistic iris textural model and adaptive thresholding. The model adaptively learns its parameters on the iris texture part and subsequently checks for iris reflections, eyelashes, and eyelids using the recursive prediction analysis. Our method obtains better accuracy with respect to the previously performed Noisy Iris Challenge Evaluation contest. It ranked first from the 97+2 alternative methods on this large colour iris database.