An Iris Verification Using Edge Detection

This paper proposes a new and fast method for iris verification. The proposed algorithm is an intensity-based approach using edge information of an iris as its signature. The edge information is detected by four different directional line detectors. Significant edges are then selected using an adaptive thresholding technique, called quotient thresholding. The merit of this thresholding is its capability in compensating changes of illumination inheriting in iris images in a database. The compensation is achieved by maintaining a ratio value between foreground and background of each image in the database. This simple technique yield informative edge features. A spatial point-to-point matching is used to measure similarity between two iris features. Our experiments illustrate promising results.