Within the biometric context, the iris is usually accepted as one of the most accurate traits particularly in access control. However, present systems demand a cooperative behavior and highly constrained imaging conditions. This restricts the range of domains where iris recognition is applied. The first step of iris recognition is a stable and robust segmentation of the iris. During the seminar you will get to know several common segmentation techniques and their application to iris images. Contributions implemented in teamwork should be submitted to the public Noisy Iris Segmentation Challenge (NICE.I). The aim of the international challenge is the evaluation of the accuracy of segmentation techniques on realistic iris images. - Requirements: 30 minutes talk, prototype implementation, 4 pages handout - For participation at the international challenge: teamwork is encouraged to develop a contribution - The topics are offered in the first session of the seminar and every student can choose one of these topics. All students who successfully participate in the seminar will be given a certificate (Seminar im Hauptstudium).

Dates & Rooms: