The Research Of Segmentation And 3-D Reconstruction In DICOM Image

With the development of computer technology, DICOM images play an important role in medical diagnosis.People have been looking for the ideal segmentation method. Three-dimensional medical image reconstruction technique is an important research topic in the field of image processing and analysis, which aims to split them from the image out of the target tissue visualization, visual results can objectively reflect the lesion site, thereby providing an objective for the clinical diagnosis according to provide a reference for doctors to develop surgical planning.According to the problems of current researches, the main work of this paper are as follows:1. Research on reading the DICOM format images.This paper proposes a method of based on ITK, VTK to read DICOM images. First,it describes the DICOM image format.Then,it describes the method of read single and multiple sequence DICOM files based on ITK, VTK toolkit. Using the two methods can be easily read and display DICOM image files and the implement of method is very simple.2. Research on the image filtering and segmentation techniques. This paper proposes a method of vertebral segmentation method based on three-dimensional surface model. First,it describes the basic concepts of mage filtering and segmentation.Then,it describes the method of programming realized threshold, median and mean filtering.After that, it propose the custom-made spine segmentation method.3.Research on three-dimensional reconstruction techniques in DICOM iamge.This paper proposes a new method of ray-cating. First,it describes the basic concepts of visualization,its common method of surface rendering and volume rendering.Then,it describes the method of programming realized the algorithm of Marching Cubes and Ray Casting. For questions of ray casting algorithm can’t meet the need of real-time interactive rendering, this paper proposes an improved method.