Doctoral Defense by Maximilian Baust

Date: Tuesday, April 27, 2012 Time: 14:00 Location: Room 03.13.010, FMI-Building, Garching, Boltzmannstr. 3

Polar Active Contours for Medical Applications

Abstract:

The accurate and robust detection and tracking of objects in different kinds of imaging data is a fundamental task in computer vision, which can be achieved by so-called active contours. The contribution of this dissertation is the development of a novel class of polar active contours, i.e. a sub-class of active contours which can only describe star-shaped objects. These novel active contours enjoy several properties, such as an increased robustness to noise for instance, which are particularly desirable for medical applications, e.g. the segmentation of abdominal aortic aneurysms from computed tomography angiography data or the tracking of the left ventricular cavity acquired by magnetic resonance imaging.