Medical Image Segmentation Using GVF Deformable Models

Over the last decade, computerized image segmentation has played an increasingly important role in medical imaging. Image segmentation is aimed at helping people to augment their understanding of various properties of objects of interest in images. Segmented images are now used routinely in a multitude of different applications, such as diagnosis, medical image vision and computer-integrated surgery. Due to both the tremendous variability of object shapes and the variation in image quality, when applying classical segmentation techniques such as edge detection and threshold, these techniques either fail completely or require some kind of post-processing step to remove invalid object boundaries in the segmentation results. To address these difficulties, deformable models have been extensively studied and widely used in medical image segmentation, with promising results. GVF deformable models are applied to extract object boundaries in this thesis and this is the base of 3D reconstruction of human tissue.The focus of this dissertation is to investigate parametric deformable models. An energy minimizing formulation is applied to get mathematical representation of an object boundary. To get a reality-based interpretation of the deformable model and its parameters, an approach originates from the theory of elasticity is proposed. The conclusion is helpful to further understand the behavior of deformable model. A new way is provided to optimize parameters.The finite difference method is applied to do numerical implementations of deformable models. Some deficiencies in the literature are improved. The strongpoint and shortage of some deformable contours is compared lively.GVF model derives from generalized force balance equations, a vector diffusion equation is employed to diffuse the gradient of an edge map in regions distant from the boundary, yielding a general type of force field which can be decomposed into an irrotational component and a solenoidal component. The excellence of GVF model is analyzed in this thesis. 2D GVF computations are implemented using MATLAB code. The image segmentations of musculature of human crus and human chest are implemented using GVF model.