New Verification Strategy for Finger-Vein Recognition System

This paper proposes a new finger-vein recognition system that uses a binary robust invariant elementary feature from accelerated segment test feature points and an adaptive thresholding strategy. Subsequently, the proposed a multi-image quality assessments (MQA) are applied to conduct a second stage verification. As oppose to other studies, the region of interest is directly identified using a range of normalized feature point area, which reduces the complexity of pre-processing. This recognition structure allows an efficient feature points matching using a robust feature and rigorous verification using the MQA process. As a result, this method not only reduces the system computation time, comparisons against former relevant studies demonstrate the superiority of the proposed method.