Fast and Efficient Iris Image Segmentation
Daniel Felix Brito,
Lee Luan Ling
Iris segmentation is one of the crucial operations involved in iris recognition. Accurate iris segmentation is fundamental for the success and precision of the subsequent feature extraction and recognition, and consequently the high performance level of the iris recognition system. Most iris segmentation approaches proposed in the literature require highly complex-exhaustive search and learning of many modeling parameters and characteristics, which prevents their effective real-time applications and makes the system highly sensitive to noise. This paper presents a fast and efficient iris segmentation methodology to address relatively simple solutions to these problems. Three major procedures involved in the proposed iris segmentation approach, namely pupil detection, limbic boundary localization, and eyelid and eyelash detection, were carefully designed in order to avoid unnecessary and redundant image processing, and most importantly, to preserve the integrity of iris texture information. The proposed iris segmentation algorithm has the following particular properties and advantages: (a) avoidance of complex geometric and mathematical modeling; (b) no need of a training phase for algorithm design and implementation; (c) guaranteeing real-time iris segmentation even for iris images with severe occlusions; (d) high accuracy in iris segmentation and therefore low segmentation error rate. Experimental results, reported in this paper, demonstrate that the proposed iris segmentation algorithm outperforms some well-known methods in both accuracy and processing speed. As a consequence, the iris recognition system that incorporates the proposed iris segmentation algorithm is capable of offering recognition performances comparable with those reported by other state-of-the-art methods.