Semi Automatic Segmentation of Articular Cartilage using Variational Methods(web-version)

Osteoarthritis (OA) is a syndrome of joint pain that affects the large weight bearing joints. It is caused by an abnormal wearing of articular cartilage, covering the joints. In addition to the inconvenience OA causes, its treatment is very time consuming and expensive. Therefore it is desirable to improve methods for an early diagnosis of OA. The detection of thinning of articular cartilage provides a good support for the diagnosis of OA in its early stage. The first step in this diagnosis process is the accurate segmentation of the cartilage surface. In this Master’s Thesis we propose an interactive segmentation framework for the semi automatic segmentation of articular cartilage. Until today, no automatic segmentation method is able achieve the accuracy, necessary for a trustworthy diagnosis. Also, physicians in general prefer to be able to control and modify the segmentation result, which is usually complicated using automatic methods. Semi automatic methods allow the user to incorporate knowledge into the segmentation process, whilst reducing the time and improving the repeatability compared to fully manual methods. The proposed segmentation model is based on a weighted Total Variation energy and minimised using efficient numerical approaches. Implemented on today’s userprogrammable graphics cards, it allows real-time user interaction. The evaluation of our segmentation method using real-world magnet resonance datasets of human knee joints shows, that we are able to speed up the segmentation process significantly, compared to manual and semi automatic segmentation methods.