3D Vessel Segmentation in CT for Augmented and Virtual Reality

Blood vessels 3D rendering has numerous applications, ranging from diagnosis to pre-procedural and surgical approaches as it enriches vessels visualization for the clinician. In this paper, we propose a 3D blood vessels segmentation method designed for use with a cone beam computed tomography (CBCT). The algorithm constitutes a module in the development of an angio-CT visualization system, based on augmented or virtual reality as instruments supporting and improving medical decisions. The proposed segmentation tool exploits a bone segmentation step for easing the extraction of blood vessels. Both steps are based on region growing technique. For each subject are used two CBCT acquisitions, where the first one is acquired with a traditional CT scan and is used for bone extraction, while the second one is acquired after contrast medium administration and is used for vessels reconstruction. This novel segmentation algorithm provides an automatic and accurate tool to segment and render blood vessels tree. 3D anatomical models were viewed through a virtual reality environment in order to validate the visualization system.