Effective Iris Recognition System

Biometric identification has provided a robust technique for proving identity and has become the standard in identity authentication and access control. Iris recognition is a prudent biometric identification system with promising results in the security systems area. In this work, our system introduces a more accurate method called RANSAC (Random Sample Consensus) for fitting ellipse around non circular iris boundaries. It can locate iris boundaries more accurately than the methods based on Hough transform. Also we used Daugman's rubber sheet model for iris normalisation and elliptic unwrapping, and correlation filter based matching for intra class and inter class distance evaluation.PSR(Peak Side Lobe Ratio) is the similarity measure used for matching templates. By these the recognition process is improved compared to Daugman's method. WVU database is used for conducting experiments and promising results are obtained.