An Efficient Approach for Simple Iris Localization and Normalization Technique
Ramadan Gad, Ayman El‐Sayed, M. Zorkany, Nawal Ahmed El-Fishawy · Menoufia Journal of Electronic Engineering Research · 2016
Iris segmentation is a critical step in iris recognition system. Some challenges to segment iris such as occlusion by eyelids, eyelashes, and corneal and specular reflection. Traditional techniques to remove these artifacts were more complex, computationally exhaustive, time consuming, sensitive to noise, and large memory occupation. In this paper, a novel approach is proposed to isolate the iris free of artifacts without changing its structure. In this method, the segmentation, normalization, and unwrapping steps are merged into one step. A series of morphological operations to remove reflections are implemented. Pupil is detected by using threshold technique. This method gives better results in factors of image quality, accuracy, and less memory occupation in general purpose processor. As a result, this proposed technique is suitable for hardware and real-time applications as it is fast, more accurate, no deformation of iris structure, and less complex.