Automated Centerline Extraction of Angiogram Vessels Based on Multiscale Gabor Filters
基于多尺度Gabor滤波的造影血管中轴线的自动提取

In this paper,a new automated approach is proposed to extract the centerlines from 2D angiogram images.The centerline extraction is the basis of 3D reconstruction of the blood vessels,so the accurate localization of centerlines counts is extremely important.The characteristics of multiscale Gabor even filters,flexible frequency bands and enhancement effects,are fully utilized to enhance centerlines of the blood vessels in various size,which greatly reduce the background noise.Then the non-maximum suppression algorithm is adopted, guided by the directions provided by calculating the Hessian matrix of the image,to get the local maximum points on the response map.Finally,vessel centerlines are extracted by a hysteresis-thresholding segmentation.The experimental results demonstrate the robustness of the algorithm and good quality of the extracted vessel centerlines.