A novel algorithm for hand vein enhancement and segmentation based on the Retinex method

摘要

The contrast is low and the venous structure is simple for the hand vein image captured by a near infrared camera. Based on the Retinex method a novel algorithm is presented for human hand dorsal vein image enhancement and segmentation in this paper. First of all, the acquisition device for human hand dorsal vein image is designed based on near infrared imaging. Secondly, the area-of-interest is automatically segmented to reduce the influence of edge and background regions. Thirdly, the Retinex Method based on adaptive filtering is used to enhance the hand dorsal vein image. Finally, the dynamic global threshold method is used to segment the hand vein image, and there are some false hand vein blocks, so the vein block determination criterion is established. The experimental results show the efficiency for the algorithm proposed in this paper.