Binary Filter for Fast Vessel Pattern Extraction

摘要

Vessel pattern extraction from images is important for many applications, such as personal recognition and medical analysis. The matched filter has been widely applied for vessel extraction because it works efficiently in enhancing vessel images. However, the matched filter is time-consuming as it uses real-value Gauss or Gabor filters to generate the matched filter response images. In this paper, we propose a binary filter for vessel pattern extraction which can achieve similar results as the Gauss or Gabor filter but with fewer parameters and faster processing speed. The proposed method can process 22 images with image resolution of 512*512 per second, which is about three times faster than Gauss or Gabor filter. Experiments on palm vein and retinal fundus images show the effectiveness and efficiency of the proposed methods.