Iris Image Identification Using Gray Level Co-occurrence Features and FMT-Based Normalization

Abstract

Iris is one of the best biometrics features for human identification and verification. In this paper, the proposed framework is consisting of iris image segmentation, normalization, feature extraction and classification. We use canny edge detector to generate an edge map and Hough transform to segment iris image. The next stage is to transform on the iris regions so that they have the same dimensions. The dimensional inconsistency between eye images are due to the stretching of the iris caused by pupil dilation from varying levels of illumination, varying imaging distance, rotation of the camera, head incline, and rotation of the eye within the eye socket. We apply Fourier Mellin Transform (FMT) to make the normalized images that are invariant to rotation and translation. Also, using Gray Level Co-occurrence Matrix (GLCM), statistical features of the normalized images are extracted. The proposed algorithm is evaluated on standard CASIA database. A multilayer perceptron classifier obtains an accuracy rate of 98.76% for iris image classification problem. The obtained results show that our proposed algorithm is more efficient than other common approach in real and operational conditions.