Fully-Automatic Branch Labelling of Voxel Vessel Structures

Today, it is possible to acquire volume representations of the vessel structures in the brain. The selfadjusting probe, a new tool introduced in a previous paper, enables semi-automatic shape extraction. The self-adjusting probe extracts shapes from a surface model. Yet, if two vessel branches are close together, it is possible that surface vertices of a neighbor vessel are included in the set of vertices used to extract the local shape of the vessel investigated. These erroneously included vertices deteriorate the accuracy of the estimated shape. Therefore, a method has been developed to give the vessel vertices a unique label per vessel branch. Now, surface vertices of neighbor vessel branches can be excluded because their label is different.