The Research On Technology Of Medical Visualization Based On Lung CT

CT scanning has been the most important method of lung disease examination. Using technology of medical visualization to process CT images, we can reconstruct high degree of visualization properties of three-dimensional (3D) images, which provide disease information to clinical diagnosis, so that disease which hard to be diagnosed in the past may be accurate diagnosed now. Technology of medical visualization contains image segmentation, 3D reconstruction, virtual clipping, 3D measurement.Through learning of region growing, mathematical morphology, curvature, rolling-ball algorithm, a lung segmentation method which optimizes pixel-based method is putted forward. Firstly, lung region is extracted from chest CT images by region growing, then smooth the border and fill holes of the lungs by mathematical morphology operator; Secondly, according prior knowledge of smoothing of border, quickly identify suspicious of the gap points on the edge of the lungs through the curvature of the threshold; Finally rolling-ball algorithm executes only near suspicious points. Through this method, time decreases, efficiency increase, and to some extent accuracy improved.In order to achieve large-scale data on the parallel visualization, a data-parallel-based solution of surface rendering is putted forward and implemented, simple and effective. The solution adopts a dynamic data distribution method and integrates complete process of ISO-surface, has excellent acceleration and effect.System function and performance testing show that, in the medical visualization system based on lung CT, the segmentation method remedy short coming of traditional methods, and has a more accuracy result. The parallel programme meets the calculation problem brought by high-resolution CT image, improves the drawing results of image, can help doctors diagnose lung disease and design lung surgery.