Blur grey-level hit-or-miss transform for fully automatic 3D segmentation of coronary arteries

International audience; 3D CT scan images of coronary arteries are complex to analyze because they provide a 3D object that is visualized through 2D projections. Medical diagnosis suffers from inter- and intra-clinician variability. Therefore, reliable software for the 3D reconstruction and labeling of the coronary tree is strongly desired. Finding appropriate methods is known to be a challenging problem because of data imperfections: noise, heterogeneous intensity... In this paper we propose a fully automatic algorithm for coronary artery extraction from X-ray data sequences of a cardiac cycle (3D-CT scan, 64 detectors, 10 phases). Our method is based on the blur grey-level HMT, and it is guided by anatomical knowledge. Our segmentation gives good result on 90% of the images, while those where it fails are very noisy. It is therefore a promising tool for the automatic 3D reconstruction of the coronary tree from 3D temporal angiographic sequences.