VDM Gouda, M: Iris Texture Analysis

An efficient method for personal recognition is iris texture analysis. This personal recognition system is considered one of the most secure personal identification systems. Here, we introduce a convenient technique for iris recognition system. First, eye image is captured using specially assembled camera system. The captured image is then ready to pass through the image enhancements stage. The next step is the iris localization, which is done using a fast and effective method that treats the illumination in the pupil. The iris image is then normalized using a new method which depends on taking sectors of the image and throwing away parts of the image occupied by the eyelids and eyelashes. The image now is ready for features extraction. Features extraction is done by the DoG filter. This 2-D filter is convolved with the normalized image giving us a 2-D real output image of size 512*64 pixels. This output is then coded, generating a binary 2-D code of size 16*32. The iris code generated is used to build the system knowledge. Matching is then done by measuring the hamming distance between two codes. This is done by applying the exclusive-or operation to both of the compared codes.