Segmentation of Abdominal Aortic Aneurysm Using Deformable Models

We present the work that has been done on segmentation of abdominal aortic aneurysm (AAA) by Image Processing Group on Faculty of Electrical Engineering and Computing in Zagreb. We investigated performance of two types of deformable models: parametric and geometric deformable models both operating on three-dimensional data sets obtained by computer tomography angiography (CTA). Both inner and outer aortic borders can be segmented. The segmentation of the inner aortic border is trouble-free due to a contrast agent inside blood flow, which results in good image contrast on the inner aortic boundary. Segmentation of the outer aortic border presents a challenge because of poor image contrast on it. Tissue and organs surrounding the aorta can have similar optical density as aortic wall and that can make border between them hard to detect. Two types of deformable models provide two different solutions to the problem. Both segmentation methods have been tested on real patient data.