3D Angiographic Image Segmentation
Benoît Naegel, Nicolas Passat, Christian Ronse · 2013
The purpose of segmentation is to help the medical expert by enabling an easy detection of specific structures. By focusing on the case of angiographic imaging (i.e. the imaging devoted to vascular structures), this chapter illustrates (non-exhaustively) the method of taking into account and formalizing anatomical knowledge elements in mathematical morphology. In particular, two concepts enabling to tackle this problem are described: the anatomical knowledge models and the hit or- miss transform. Two examples of segmentation methods based on these concepts are then presented, in order to emphasize the soundness of the proposed approach. In order to process spatial information contained in medical images, the use of robust anatomical hypotheses is essential. The hit-or-miss transform is an operator enabling the detection of some configurations of points satisfying constraints related to two structuring elements. Controlled Vocabulary Terms angiocardiography; image segmentation; mathematical morphology