Iris Recognition Using Segmental Euclidean Distances

A novel segmentation method is developed for segmenting an Iris image into 6 circular segments after it has been cropped from an eye. Each segment is treated as a separate image to find Eigeniris and to recognize the Iris by the segmental Euclidean distance between the features of training and test samples computed from their segments. In addition to the Eigen features, the DCT and fuzzy features are extracted from the segments of iris images. Instead of segmental Euclidean distance as a classifier, the Support Vector Machine is also employed for the classification. The evaluation of performance on the CASIA V.2 database using Eigen, DCT and fuzzy features has resulted in the recognition rates of 93%, 98% and 95% respectively.