Selective Detail Enhancement Algorithm for Finger Vein Images

Abstract

AbstractIn near-infrared finger vein images, the contrast between vein texture and skin is low, and traditional algorithms based on enhancement filters and grayscale stretching are difficult to effectively enhance the vein structure, and it is easy to introduce noise. Therefore, a selective enhancement algorithm for finger vein images based on adaptive guided filtering is proposed. Firstly, a gradient operator weighted guided filter with better edge-preserving ability is proposed to decompose the vein structure of finger vein images, and obtain smooth layer and detail layer. Secondly, a selective weighting factor is designed for vein detail layer enhancement while reducing noise interference. Finally, the enhanced detail layer is fused with the smooth layer to obtain an enhanced image of finger veins. The experimental results show that the proposed algorithm can effectively improve the quality of finger vein images. In six popular image quality evaluation indicators such as information entropy (IE) and peak signal-to-noise ratio, the image recognition rate after enhancement is also improved due to mainstream algorithms.KeywordsFingervein enhancementGradient operatorGuided filtering