For her Phd project, Dieuwertje works on finding parameters associated with rupture of abdominal aortic aneurysms (AAA’s), based on image data. To analyse an aneurysm, we need a watertight 3D model of the aorta. Explicit representations of the vessel geometry are challenging to acquire directly from image data. At the STACOM workshop, Dieuwertje presented an AI-based method to acquire a vascular model based on a set of 3D points on the surface.
The key idea was to represent the surface by an implicit function, rather than explicitly. Therefore using a neural network to represent this function, that was optimized based on the set of points on the surface. Some benefits of these function representations: it captures anatomical priors and allows for smooth blending of vessels.
We are happy to announce that this work was awarded with a best poster award at the STACOM workshop at MICCAI 2022 in Singapore.