IDENTIFICATION OF THE RACE LIVELINESS AND MATCHING OF THE IRIS IMAGES

Abstract

Individual recognition using Iris is most commonly employed in the entire place. This  requires some special cameras to take the Iris image and the obtained iris images were classified based on the features extracted. Only identification of person is not the application of Iris recognition. It can be developed to do many processes. In this process the race of the person is identified. Also the input iris is fake or original is also identified. The iris image captured from the cameras was taken. Some Iris recognition system fails because some of the recognition systems identify the iris image in photos; also some of the recognition systems 
identify the fake iris images generated by some attackers. An iris recognition system that identifies the fake iris images are much needed. The proposed
iris recognition system identifies the fake and original iris images. In order to identify fake and original iris image SIFT descriptors were derived from the input iris images.  Before this process the iris image is preprocessed by resizing the iris image and then filtering the iris image using median filter. The preprocessed iris image is then normalized. The normalization process identifies the iris and pupil region in the image correctly and it reshapes the identified positions. The normalization process improves the efficiency.The SIFT descriptors identifies the corner points in the iris image and then the extracted points were used to generate the codebook. The code book generation processes is done using Hierarchical Visual Code book Generation method. As performance metrics we have taken accuracy, sensitivity and specificity. The calculated values shows that the proposed system has better accuracy and also additional methods are also included compared to the existing methodologies.