Medical image segmentation using edge-based active contours.

The main purpose of image segmentation using active contours is to extract the object of interest in images based on textural or boundary information. Active contour methods have been widely used in image segmentation applications due to their good boundary detection accuracy. In the context of medical image segmentation, weak edges and inhomogeneities remain important issues that may limit the accuracy of any segmentation method formulated using active contour models. This thesis develops new methods for segmentation of medical images based on the active contour models. Three different approaches are pursued: The first chapter proposes a novel external force that integrates gradient vector flow (GVF) field forces and balloon forces based on a weighting factor computed according to local image features. The proposed external force reduces noise sensitivity, improves performance over weak edges and allows initialization with a single manually selected point. The next chapter proposes a level set method that is based on the minimization of an objective energy functional whose energy terms are weighted according to their relative importance in detecting boundaries. This relative importance is computed based on local edge features collected from the adjacent region inside and outside of the evolving contour. The local edge features employed are the edge intensity and the degree of alignment between the images gradient vector flow field and the evolving contours normal. Finally, chapter 5 presents a framework that is capable of segmenting the cytoplasm of each individual cell and can address the problem of segmenting overlapping cervical cells using edge-based active contours. The main goal of our methodology is to provide significantly fully segmented cells with high accuracy segmentation results. All of the proposed methods are then evaluated for segmentation of various regions in real MRI and CT slices, X-ray images and cervical cell images. Evaluation results show that the proposed method leads to more accurate boundary detection results than other edge-based active contour methods (snake and level-set), particularly around weak edges