基于双源CT的心脏虚拟内窥镜显示-医学图像处理专业论文.docxVIP

II II 最后，用前面分割得到的图像，获取轮廓并得到心脏内部结构的 轮廓线，实现三维面重建，即虚拟内窥镜。 研究结果表明，本文的自适应种子填充法对于大数据量的分割非 常有效，Contourlet 变换进行任意方向角度的线状结构的分割和识别 也是非常有效的，而虚拟内窥镜的实现对于医生进行心脏结构的观察 是有一定的临床价值的。 关键词：双源 CT, 自适应种子填充法, Contourlet 变换, 图像分割, 虚拟内窥镜,三维面重建 III III CARDIAC VIRTUAL ENDOSCOPY DISPLAY BASED ON DS CT IMAGES ABSTRACT Virtual Endoscopy is an application of virtual reality widely used in modern medicine. It simulates optics endoscopy by use of medical images and incorporates Image Processing, Computer Graphics, Computer Vision, Virtual Reality Technology, etc. It is a complete untouched diagnosing method that can avoid using actual optical endoscopy camera invading into human body. The aim of doing research of virtual endoscopy is to provide trusty means of medical diagnoses, and also use it in new application fields as assistant diagnoses, surgery planning and training. Heart is one of the most important organs for human. Through pumping function of heart, blood circulates throughout the whole human body and provides energy and oxygen to the other organs of human body. Mitral valve malfunction is one of the most common diseases of the heart. Mitral regurgitation (MR) is usually the result of close imperfect of mitral valve, and leads to blood supply insufficiency. So simulate real environment of inner-heart and observe mitral valve and its movement clearly is very important for doctor’s clinical care. In this paper, we try to segment region of interest of human heart that in the contrast medium and segment mitral valve. At last achieve virtual endoscopy through the segmented images. At first, we enhance and segment the region of interest of human heart based on DS-CT imagers. Consider the large amount of images in this case, we propose adaptive seed-filling algorithm to solve this problem. Then, Contourlet transform is used to enhance and segment heart valve based on the part- segmented images. At last, we get the contour of all these images segmented and realize PAGE PAGE IV 3-D surface rec