Automatic Parameterisation and Registration of Abdominal Aortic Aneurysm

We propose a fully automated pipeline for analyzing aortic abdominal aneurysms starting from CT scans. The initial step of the pipeline is retrieving segmentations of patient aortas from a dataset and isolating the region of interest. The geometries are parameterised first computing the centerline of each segmentation. Then the centerlines are used as a reference to approximate the aneurysm wall, thus representing the shape of the aneurysm with a simple mathematical model. Subsequently, orthogonal planes are defined based on the cylindrical topology of the aneurysm, and sections of the aneurysm are identified. The main vessel in each cutting plane is recognized, and the sections are approximated in Fourier space. The final steps involve rigid registration of the parameterised geometry and the calculation of displacement maps, enabling quantitative and rigorous comparison of aneurysm features and properties across multiple images.