الفهرس | Only 14 pages are availabe for public view |
Abstract Iris recognition systems capture an Image from an individual’s eye. The iris in the image is then preprocessed (segmented and normalized). After that unique and robust features of each iris• are extracted and used to match with iris patterns in an enrolled database. The performance of iris recognition systems highly depends on the segmentation process. For instance, even an effective feature extraction process would not be able to obtain useful information from an iri.;; image which is not segmented properly. The segmentation process is used to isolate the iris from the surrounding components as the sclera, pupil, eyelids, and eyelashes where detecting the iris inner and outer boundaries at the pupil and sciera. detecting the upper and lower eyelid boundaries if they occlude, and detecting and excluding any superimposed eyelashes or reflections. In noncooperative iris recognition one should deal with uncontrolled behavior of the subject as well as uncontrolled lighting conditions. That means eyelids and eyelashes occlusion, non uniform intensities, reflections, imperfect focus, and orientation among others are to be considered. To cope with this situation a noncooperative iris segmentation algorithm based on numerically stable direct least squares fitting of ellipses model and modified Chan- Vese model (local binary fitting energy) with variational level set fonnulation is to be proposed. The proposed algorithm is tested using CASTA-Iris V3. Securing information and ensUrI;lg the pnvacy of personal identities is a growmg concern in recognition systems. Biometrics¬based authentication schemes have usability advantages over traditional password-based authentication schemes. Ho\vever, biometrics raises several privacy concerns, it has disadvantages comparing to traditional pass\vord in which it is not secured and non revocable. We propose a fast method for securing revocable iris template using parallel-bit stream \vatennarking to overcome these problems. Experimental results prove that the proposed method has low computation time and is available to protect iris template and enhance the security of the iris recognition system. Keywords: noncooperative iris segmentation, nwnerically stable DLS~ modified Chan- Vese model, security’, revocable, watermarking, bit stream. |