Iris Recognition System based on ZM, GF, VR and Matching Level Fusion

Isis is the physiological biometric trait used to recognized a person efficiently. In this paper, we propose Iris Recognition System based on ZM, GF, VR and Matching Level Fusion. The Region of Interest (ROI) of iris is extracted using segmentation. Zernike Moments (ZM) is applied on segmented iris images to extract ZM features. The novel concept of many feature vectors of a single person are converted into single vector per person ie., Vectors Reduction (VR). The Euclidian Distance (ED) is used to compare feature vectors in the database with feature vectors in test section to compute the performance parameters. The Gabor Filter (GF) is also used to extract features of iris. Many GF feature vectors of single person are connected into single feature vector per person. The ED is used to compare database and test feature vectors to compute performance parameters. The performance parameters obtained from ZM and GF are fused using normalization technique to improve the performance parameters. It is observed that, the performance parameters are better compared to existing techniques