Study Of Medical Image Registration And Fusion Using 3D Contour-Surface-Based Method

The imaging theory of all kind of medical image are difference. Each medical image show different body's information from different point of view.We can show all kind of images fall together with the help of information fusion technology based on medical image registration. Fusion image can show all kind of information of body in only one image, which making the best use of the advantage of each image modality. And thus the fusion image can show all status of the inner structure and function and give information of body anatomy, physiology and pathology directly. Physician can make the best use of the medical image gained from many modalities and diagnose disease from many angle of view.The main purpose of this paper is to study the image registration of CT and MRI image with the image registration algorithm of 3D contour-surface-based which belong to intrinsic registration methods, one of landmark based. According to this arithmetic, it carries through the image segmentation first, pick-up the contour curve or object's surface ,thus it simplifies the registration of 2D or 3D image to the matching of 2D curve or 3D surface and don't consider the inside pixels of image.All images of CT and MRI adopted in this paper are standard Dicom3.0 format, which is introduced briefly. Then the reading and saving of Dicom3.0 file are realized by programming. Then, we carry through the image segmentation by means of 8-connected boundary search method in the boundary tracking method, search the boundary of image, obtain the contour surface of CT and MRI image of patient, set up the point set of contour. Next, 3D reconstruction of the CT and MRI image are done separately by the point set of contour, forming the 3D surface model of contour. The next key step is to register the CT and MRI contour surface according the point set of contour. We define the mutual distance between two contours of two image modalities, calculate the mutual distance between CT and MRI contour and optimize the mutual distance as the comparability measurement of the registration. Before we...