Study On The Techniques For High-resolution3D Visualization Of Pulmonary CT Images

The high-resolution three-dimensional (3D) visualization of anatomic structure inmedical images is a key component for computer-aided diagnosis (CAD), image-guidedsurgery and therapy (IGST), where rendering what the radiologists or surgeons highly concernwith in images in clear and verisimilar fashion. Due to the high-resolution three-dimensionalvisualization of anatomic structure can provide three-dimensional characteristics and thespatial location of it, relatively easy to display anatomical and pathological changes, helps toform the morphology characterization (judging benign and malignant nodules base on thesurface texture feature, etc.) and surgical navigation simulation system design, so thistechnology has attracted much attention.But the organizational structure in human body is very complex and data quantity of thepulmonary organ CT sequence images is large. In order to be able to carry on3D rendering ofit an standard PC environment, to reduce the overall amount of data is needed, which isnegative for the doctors’ interested smaller groups (such as pulmonary nodules) rendering. Sothe study of local high-definition display and the overall3D rendering is the content of thisstudy. The main research work and achievements are as follows:(1) Using a variety of segmentation method for complex structures in pulmonary CTimages, depending on the organizational characteristics selected segmentation method andorder, and repair flaws or concavity along the boundary of segmented region due to inter-pixelgray value discontinuity.(2) According to the medical image sequences characteristics, applying the single or thetwo-dimensional images POCS method in a3D improvement. In the motion estimationprocess, for the spatial characteristics of three-dimensional volume data, designing the motionestimation process equation, and estimate for volume data matrix are obtained. Experimentalresults show that the new reconstruction algorithm can improve the spatial resolution of thevolume data while maintaining the similarity with the original data, the recovery effect isbetter than traditional interpolation method. (3) Integration characteristics of two kinds of rendering method, a hybrid renderingmethod for different anatomic structure segment results is presented. Surface renderingtechniques can get a positioning accuracy clarity surface, the correct topology structure ofthree-dimensional model; volume rendering method can get integral effect of all voxels alongthe observed line of sight and remain necessary details of other anatomic structures.Experiments show that the hybrid method can visualize on a standard PC with completesequence of images, and is efficient and feasible in comparative with the visualization ofunitary resolution or single rendering method.