Finger Vein Feature Extraction Based on Linear Weighting Function Immune Clone Algorithm

Abstract

To solve the misjudgment of noise and vein information in features extraction from low quality images, a novel method based on LWF (Linear weighting function) immune-clone algorithm is proposed in this paper. The method can produce initial antibody by using adaptive threshold method, obtain weighting function by curve fitting, and denoise and enhance border by linear weighting of the vein area. The function of affinity and concentration of antibodies helps to boost the growth of the vein information and suppress the interference of noise. Simulation results show that compared to other algorithms, finger vein pattern extracted by the algorithm proposed in this paper is more distinct and accurate. In addition, this algorithm, which can effectively retain the details of information, is especially suitable for features extraction from low quality finger vein images.