Fusion of Local and Global Iris Features to Construct Feature Vector Using Genetic Algorithm

Extracting important features from an image is a complicated phase in the field of image processing, biometrics and computer vision. After crossing over phases such as denoising, segmentation and normalization in any pattern (Iris) Recognition System, feature extraction takes place which represents the features in the form of numerals or binary known as feature vector. In Iris Recognition System (IRS), this feature vector is named as Iris Code which determines the matching or similarity score when test image is compared with database image for verification/authentication of an individual. In more general, IRS has used wavelet transform for feature extraction. In this paper, a new feature extraction technique is proposed that fuses both local and global properties of a normalized iris image using two cross over scheme in genetic algorithm and constructs 64 bit binary feature vector. The proposed technique is experimented using MATLAB 12a and the execution time for the proposed system is calculated. Based on the experimental results it is clear that the proposed feature extraction technique frames the Iris Code with success that minimizes an elapsed time to 196.7ms. And also it gives 98.75% accuracy.