Skip to main content

Super-Resolution for Iris Recognition

Researcher: Vijayakumar Bhagavatula

Research Area: Security of Cyber-Physical Systems

Abstract

Scope: Iris recognition is known to offer excellent identification performance among all biometric modalities --- but this is true under highly controlled scenarios where the iris images are of high quality. In real-world applications, the image quality degrades due to several factors, e.g., the subject being far away from the camera (i.e., low resolution) and subject not looking at the camera. The goal of this project is to develop and evaluate superresolution algorithms aimed at recognition of iris images --- in contrast to existing super-resolution approaches that are aimed at image reconstruction. Super-resolution recognition will be achieved by explicitly including iris recognition features in the super-resolution optimization formulation. This approach will be evaluated using low-res versions of high-quality iris databases (e.g., University of Notre Dame iris database) as well as challenging iris image databases (e.g., UBIRIS from Portugal).

Outcomes: New super-resolution algorithms for recognition of low quality iris images. Conference and journal submissions summarizing the super-resolution for iris recognition algorithms and results. A proposal to be submitted to IARPA or other appropriate funding agencies .