Abstract: In this paper, a novel segmentation algorithm is presented to improve segmentation performance for vein images. The proposed algorithm uses the distribution ratio of directional fields (DRDF) as the segmentation principle to distinguish vein lines from background. First, the neighbor information and template of neighbor segmentation are used to obtain directional fields for every pixel in the original image. Then in image of directional fields, DRDF of complementary semi-circle regions and distribution discrimination function are used to calculate the 8-grayscale image. Finally, a threshold is set to binarize the 8-grayscale image to obtain segmentation results. The algorithm combines the feature of vein image and spatial properties of directional fields image. Experiment results show that it is very effective for vein image with non-uniform illumination, inhomogeneous thickness, and blurry boundary.