Implementation of iris occlusion estimation methods

This proposal presents an approach for an iris identification method for checking the identity of person using the unique nature of iris. The recognition system consist of segmentation stage which uses Hough transform and it will locate the circular iris and pupil area, occluding eyelids and eyelashes, reflections. The segmented iris area is normalized into rectangular block of fixed dimension.

Phase, amplitude and surf descriptor points are used as three different features which uses Gabor filter and Surf algorithm for feature extraction purpose. The Hamming distance, Euclidean distance and Pearson-correlation test are used for matching stage of test and referenced image.

Share This Post