Vein biometric security using irreversible curve fitting accounting for minimum storage

Abstract

Vein biometric is hidden so provides more secrecy as compared to other biometric traits. The paper presents a novel approach in vein biometric system that accounts for minimum data storage but still keeping up with maximum information. A new strategy is presented for template generation and storage. Polynomial curve fitting process has been applied on thinned vein pattern to output most uniform best approximation vein curves. Matching has been done on basis of Hausdorff distance. Obtained experimental results are encouraging and have been claiming 0% FAR and 0% FRR. The amount of storage needed for template data has been reduced greatly. The data stored is irreversible to generate the template again even when no template encryption scheme has been used that provides attack resistance to the system.

Vein Biometric Template Security Using Overlapped Shares Accounting for Minimum Storage

“…3. The matching is done using Hausdorff distance [10] which computes similarity between two images A and B by computing distance from A to B and vice versa and then considering among 2.…”

“…Curves below that length are considered to be noise. If the distance between two template data is greater than matching threshold, then the query may be termed as a match [10,13].…”