Iris Localization Scheme Based on Morphology and Gaussian Filtering

One of the basic techniques in iris recognition system is iris localization. To find a fast, effective and exact iris localization algorithm is the key step of iris recognition. After analyzing the principle, strong points and short points of some common used iris localization methods, a morphological theory based iris localization algorithm was proposed in this paper. Based on the concept of all the object can be regarded as the subset in Euclidean space, and this subset can totally reflect the shape, volume, texture and grey value of the object, so we can apply the morphology operation to identify the feature in different eye area. The proposed iris localization algorithm is as follow: iris image preprocessing at first, where we apply a suitable Gauss filter to lessen the influence of noise. Then apply the morphology operation to extract the inner and outer iris edge, determine the iris area. The proposed algorithm was tested using CASIA iris database(V1.0). And the experimental result shown the method was superior in processing speed under the same localization precision.