Extrakce objektů z medicínských obrazů na základě metody aktivních kontur

Abstract

This bachelor thesis is focused on the articular cartilage extraction from the knee image records acquired by the magnetic resonance (MR). A semi-automatic method based on the active contour is used for the segmentation purposes. The segmentation algorithm is implemented in the SW MATLAB. In the thesis, an effectivity of basic image preprocessing methods is compared with the aim of improving the segmentation results by the active contour model. The consequent extraction of articular cartilage is based on the binarization of the color map which represents one of the segmentation results. The extracted area is consequently validated on the base of the manual segmentation. One of the issue is the validation weather the active contour method is able to detect the early articular cartilage pathological changes. The method has ambitions to be used in the clinical practice in volumetric measurements, or 3D articular cartilage models.