Geometric Deformable Models Analysis And Research On Its Application In Medical Image Segmentation

Medical image segmentation is one of the essential step of medical image processing, and it plays a crucial role in both biomedicine research and virtual surgery applications such as study of anatomical structure, quantification of tissue volumes, localization of pathology, diagnosis, treatment planning, and computer aided surgery, etc. As a result, accurate segmentation method is crucial to the follow-up analysis.This paper aims to do some applicational simulation and algorithm improvement research on medical image segmentation algorithms. Based on analyzing geometric active contour characteristic and comparing advantages and shortcomings of various deformable model segmentation algorithms, we mainly study on the feasibility and improvement approaches of applying Level Set deformable models and on the basis of Level Set theory, application of Fast Marching method to medical images is studied.Firstly, in this work, basing on Level Set method and combining with the contour energy conception of Snake deformable model, the first modification integrates the average energy of the whole advancing front in traditional Fast Marching method. Then add to incorporate the gray level information of the target region into the speed term to let the evolution curve advance in the target region to solve the "boundary leaking" problem of the traditional Fast Marching method.Secondly, the improved segmentation algorithm combining Fast Marching and Level Set method is proposed in order to use Level Set method's advantages. In the simulational experiment, we can point multi-seed to extract the desired boundary of the hole in objective image. The results show that this method can remove the small regions obtained from Fast Marching method and converge the desired boundary.In the last, the paper introduces the Mumford-Shah model and C-V model. Our model has a Level Set formulation, interior contour are automatically detected, and the initial curve can be anywhere in the image. We have successfully applied this model to medical image segmentation and implemented the multi-phase Level Set model. The result show that this model can overcome the shortage of the classical segmentation model by using the global information of image to make curve stop at the edge of the object, and can detect objects with very smooth boundary or even with discontinuous boundaries.This research integrates with the virtual surgery simulation project. The research production was applied in virtual surgery three-dimensional re-construction, which shows good research value and application level.