Asian & Non-Asian Iris Image Optimum Classifier Using Generalized Feed Forward Neural Network

Abstract

Analysis of iris pattern has been the focused of researches, mostly in biometric authentication. The researchers state that the iris contains unique patterns to each person and stable with age. Therefore, iris has been used as recognition of a person. Iris recognition is one of the most reliable methods for identification of individuals. However, there are many parameters which can attenuate the accuracy and reliability of iris recognition systems. The iris recognition system’s reliability can be challenged and the accuracy of biometrics recognition systems can be degraded due to the several different abnormalities in the pattern of iris tissue. The main purpose of is to find appearance primitives of iris images firstly, compact and yet discriminative visual features, we call them Iris-textons here, are automatically learned from a set of training images. Main purpose of this work images are classified into two ethnic categories, Asian and non-Asian human iris images. Classification of Asian and non-Asian iris images is an essential research topic as it may be advantageous in monitoring biometric authentication. Therefore the need for fast, automatic, less expensive and accurate method to classify Asian and non-Asian iris images is of great realistic significance.