Research On Medical Image Segmentation And 3D Visualization

This thesis mostly studies the technologies of segmentation and three-dimension visualization of medical images. The visualization technology is an important part of image processing and analysis, which can make up the shortage of imaging equipments and provide perfect three-dimension image for users. Doctors can view and analyze from multi-angle and multi-hierarchy, therefore they can analyze and process the medical data effectively. The visualization plays an important role on assistance diagnosis, surgery simulation, guided treatment and so on. One main task of visualization is to segment the interesting region from the surroundings by analyze characters of described object and calculate the information of dissection, pathology and physiology in segmented region. Hence, researches on image segmentation and visualization play an important part in academy and have wide applications.Medical image segmentation is the basis of after treatment such as three-dimension visualization and mock surgery; it is of great importance in image processing and analysis. This thesis gives a research on the application of region growing and fuzzy connectedness in medical images. On the basis of advantages and disadvantages of them, this thesis proposes a new segmentation method based on region growing and fuzzy connectedness, which combined with anisotropic diffusion and canny edge detection, and which segments the interesting region of medical images accurately. The experiment results prove the new method is quite effective when the destination object has fuzzy edges.There are two main technologies in three-dimensional visualization, one is surface rendering, the other is volume rendering. In surface rendering aspect, this thesis emphasizes on Marching Cubes (MC) method which is very popular now, and introduces the principle of MC and the solved method of its ambiguity specifically. It gets a good reconstruction result of human slice images by using MC algorithm which is implemented in Windows system and used C++ programming language. At the same time, this thesis gets a detailed analysis of reconstruction result and proposes improved methods. At last, according to the new segmentation algorithm, this thesis implements surface rendering based on segmentation to solve the problem of traditional MC method can't reconstruct the object whose edge is fuzzy, which firstly segments the interesting region from medical images and then reconstructs using visualization technology.In the aspect of volume rendering, this thesis emphasizes on the basic principle and implementation procedure of Ray Casting algorithm, it also discusses the key technology of Ray Casting algorithm such as illumination model, picture composition and settings of transfer function. The Ray Casting algorithm is implemented using C++ programming language. On this basis, this thesis gives a deep research on three-dimensional interactive incision, which discusses plane incision and bounding box incision in detail. The experiment results prove they are effective and practical.Finally, according to the segmentation algorithm and visualization algorithms introduced, this thesis designs and implements a medical image processing system. In this system, the medical image toolkits, ITK and VTK, are adopted. Image segmentation, image registration, image visualization and three-dimensional interactive incision are implemented in this system. The user interface of this system is friendly and operated easily.