Efficient iris recognition system based on contourlet and gabor features

Biometrics based security applications are one of the evergreen research areas. Irrespective of the presence of several biometrics, iris sets mark owing to its permanence and constancy. Understanding this fact, this work proposes to employ iris as the barricade to access services. When a iris image is passed on to the proposed iris recognition system, the iris image is segmented by integro-differential operator and the contrast of the image is enhanced by histogram equalization technique. The contrast enhanced image is then normalised by Daugman’s rubber sheet model and the contourlet, gabor features are extracted. Finally, Extreme Learning Machine (ELM) is employed to match the test iris image with the trained feature set. When the images match with each other, the access is granted else denied. The performance of the proposed approach is tested in terms of accuracy, sensitivity, specificity and time consumption over four iris databases. However, the performance of the proposed approach is stable and consistent.