Three-dimensional Segmentation Of Medical Image

With the development of medical imaging technology, a variety of different modes of medical imaging technology are widely used in clinical observation of the disease and the diagnosis. This can help the doctor the condition of the patient directly and clearly understand. The interest areas of Diagnosed the illness are extracted from medical imaging by using segmentation algorithm. In the post-processing, the use of computer display technology can be more intuitive on analyzing the severity of the lesions, and quantitative measurements. In the thesis, take the vascular the human vital organs as the segmentation object, this view of the above issues the following work:Firstly, the paper describes the background of the medical vascular image segmentation and analyses some problems may encounter, it is helpful to explore and clear the target of vessels segmentation.Secondly, in order to solve the difficulties of vascular segmentation, the paper studies a variety of image segmentation, and analyses the advantages and disadvantages of various algorithms. Especially, we take more time to introduce region growing and the level set method in detail. According to the characteristics of each algorithm, this paper combines rough segmentation with fine segmentation method to segment vascular image. In the rough segmentation step, we used region growing segmentation to guide image segmentation; and in fine segmentation step, application of level set segmentation theoretical to process and guide fine segmentation.Thirdly, in rough segmentation step, we present a novel segmentation technique which is based on Grey relational analysis of Grey system theory. According to global and local inter-pixel correlation, we can extract the interest region from original images. The experimental results show that our algorithms can effectively extract the target area, and has strong anti-noise capability,what's more, this new method is simple, fast, and works well for some images. In the fine segmentation stage, combine with the results after rough segmentation and use the narrow band to optimize the speed level set algorithm .In this paper, we construct a model of limit narrow band model and set a new stop criterion, which greatly accelerated the speed level set segmentation algorithm.Finally, we propose the difficulties of three-dimensional image segmentation, and introduce the three-dimensional image segmentation preprocessing step. Based on the research of two-dimensional image segmentation algorithm, optimize the rough segmentation and fine segmentation process reasonably, and apply the complex algorithm to the three-dimensional image segmentation. And we put forward directions and goals in the future.