Iris feature extraction using gabor filter

Biometric technology uses human characteristics for their reliable identification. Iris recognition is a biometric technology that utilizes iris for human identification. The human iris contains very discriminating features and hence provides the accurate authentication of persons. To extract the discriminating iris features, different methods have been used in the past. In this work, gabor filter is applied on iris images in two different ways. Firstly, it is applied on the entire image at once and unique features are extracted from the image. Secondly, it is used to capture local information from the image, which is then combined to create global features. A comparison of results is presented using different number of filter banks containing 15, 20, 25, 30 and 35 filters. A number of experiments are performed using CASIA version 1 iris database. By comparing the output feature vectors using hamming distance, it is found that the best accuracy of 99.16% is achieved after capturing the local information from the iris images.