Extracting sclera features for cancelable identity verification

In this paper, we propose a novel sclera template generation, manipulation, and matching scheme for cancelable identity verification. Essentially, a region indicator matrix is generated based on an angular grid reference frame. For binary feature template generation, a random matrix and a local binary patterns (LBP) operator are utilized. Subsequently, the template is manipulated by user-specific random sequence attachment and bit shifting. Finally, matching is performed by a normalized Hamming distance comparison. Some experimental results on UBIRIS v1 database are included with discussion.