A method for giant aneurysm segmentation using Euler's elastica

Abstract Computed Tomography Angiography (CTA) is a medical modality having the advantage to reveal anatomical accuracy essential for delivering a precise diagnosis and an appropriate patient’s management for treating a cerebral aneurysm. Segmentation of aneurysm shapes through CTA medical images is an important step for geometrical quantification and assessment of rupture risk of cerebral aneurysms. Despite intensive researches in image processing applied to image sequences, aneurysm segmentation remains a major challenge that depends on the geometry and positioning of the aneurysm into the brain. In this paper, the segmentation analysis is performed by using GAC and Euler’s elastica models. With Geodesic Active Contour model (GAC) giant aneurysms (Size > 25 mm) are segmented in high-contrast region, while Euler’s elastica model estimates the missing boundaries caused for instance by the low contrast of a thrombus (clot within the cavity during endovascular aneurysm repair) with respect to the neighboring tissues. The experiments indicate that our original model is relevant for segmenting the aneurysm cavity and thrombus both with clear and distinct edges well providing a valuable tool for biological mechanisms understanding such as thrombosis into giant aneurysms, and medical validation.