Vein Biometric Template Security Using Overlapped Shares Accounting for Minimum Storage

Abstract

A human vasculature-based algorithm has been proposed for template security. The complete biometric template is binarized and divided into specific number of shares. These shares are overlapped so that all the binary bits are overlapped. This leads to formation of entirely new pattern. The vein biometric information is kept secured in the overlapped shares but in an encrypted way. These overlapped shares are stored as coefficients of the vein curves instead of image for individuals. This exponentially reduces space to store the templates. If anyone steals this secured template then also no useful information regarding the vein biometric can be obtained. The proposed method has been implemented and proved for very small false acceptance and false rejections on standard datasets.