A diffusion optical model for skin scattering suppression in finger vein image restoration

Abstract

Human identification using finger vein has been adopted in number of fields due to its high accuracy of recognition. However, image degradation caused by the strong scattering of finger tissues inevitably reduces the recognition rate. In this paper, we propose a new method to address the skin scattering problems in finger vein imaging. Firstly, a novel diffusion Point Spread Function (diffusion-PSF) model is proposed to precisely describe light scattering in finger tissues. Secondly, the blur-SURE (Stein's unbiased risk estimate) method is utilized to yield accurate estimation of the diffusion-PSF model's parameters. Finally, we adopt the multi-Wiener SURE-LET (Linear expansion thresholds) approach to improve the robustness of restoration performance. The experimental results illustrate that the proposed method significantly improves the clarity of finger vein images and enhances the venous network.