【作者】 周礼平；

【导师】 高新波； 梁继民；

【作者基本信息】 西安电子科技大学 ， 电路与系统， 2004， 硕士

【摘要】 虚拟内窥镜技术是当前虚拟可视化研究的一个重点，是计算机技术、计算机图形学、计算机图像处理以及虚拟现实技术应用于医学内窥镜系统而形成的一种独特的技术。虚拟内窥镜是近几年出现的医学图像高级后处理系统，由于它是一种非侵入型、无接触式的零创伤医疗检查手段，使得它在诊断医学和解剖教学领域中的应用越来越广泛。 本文针对虚拟内窥镜系统的关键技术和实现方法进行了系统深入的研究。首先，研究了医学图像的分割标注技术，提出了一种感兴趣区域(ROI)的快速模糊分割算法，在保持高精度的情况下实现了ROI的快速分割；其次，针对医学图像体数据边缘模糊的问题，提出了一种基于二维直方图的交互式三维边缘增强算法，使得体数据可以进行交互式边缘增强，方便用户跟踪分析感兴趣的内容；在虚拟内窥镜系统的研究中，提出了一种改进的基于区域增长技术的自动漫游路径提取算法；同时针对表面重建的伪影问题，提出了一种基于各向异性的平滑算法，使得内窥效果更加逼真；最后，通过分析虚拟内窥镜和体数据可视化方法的特点，提出了一种基于混合体绘制的虚拟内窥镜算法，进一步提高了虚拟内窥镜系统的逼真效果。更多还原

【Abstract】 Virtual endoscopy is a hot research topic in the field of scientific visualization, which unique technique is an important application of computer, computer graphics, image processing and virtual reality in endoscopy system. It is a advanced medical image post-processing system emerging only several years ago. Since it is a non-invaded, non-touched means and non-hurt examination method, virtual endoscopy is widely used in diagnostic and teaching in anatomy recently.Some key techniques and realizations of virtual endoscopy are systematically investigated in this thesis. Firstly, for the medical image segmentation problem, a fast fuzzy c-means (FCM) segmentation algorithm is proposed for the region-of-interest (ROI). Secondly, an interactive 2D histogram segmentation algorithm for interactively segmenting 3D data is also proposed. With this fast interactive segmentation method, scientists can study the interesting parts conveniently. Thirdly, the interactive virtual endoscopy techniques are investigated, and an automatics wandering fashion in virtual endoscopy system implemented. For the false contour problem of the surface rendering, an anisotropic diffusion smoothing algorithm is presented for surface blur. Finally, by analyzing the characteristic virtual endoscopy and volume data virtualization, an integrated volume virtual endoscopy algorithm is proposed to further improve the vivid effect of the virtual endoscopy system.更多还原