Semi-Automatic Segmentation of Coronary Arteries in CT Images
Sammanfattning: Coronary heart diseases is one of the biggest health problems in the world today. By segmenting the coronary arteries in medical images and examining them, important information about abnormal narrowing and plaque, which are main causes to these diseases, can be found. Manual segmentation of the coronary arteries are time consuming and dependent on the observer, which makes the need of automatic segmentation techniques apparent. The aim of the thesis is to implement an accurate and time-efficient algorithm to segment coronary arteries in computed tomography (CT) images. To do this, a model based algorithm combined with a multiple hypothesis approach has been implemented. This was first done in 2D and tested on manually made phantoms. Later on the algorithm was expanded to 3D, tested on phantoms and also on CT images of human hearts. The thesis has been performed for the company Medviso. Medviso has created a software for cardiovascular image analysis, called Segment. All the implementation in this thesis has been preformed in Segment. The algorithm was validated using two different datasets obtained from the Rotterdam Coronary Artery Evaluation Framework. Results from these show that the proposed method can be used to segment coronary arteries and that it, using only one user interaction, on average finds 64% of the sought for vessels with a tracking accuracy close to manual delineation.
HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)