医学断层图像快速可视化技术的研究与应用

【作者】 周敏；

【导师】 黄芳；

【作者基本信息】 中南大学 ， 计算机科学与技术， 2009， 硕士

【摘要】 医学断层图像三维可视化是可视化计算中的一个重要应用领域。现代CT医疗影像设备所获得的医学体数据量越来越多,使得三维可视化所处理的数据量巨大,因此,在保证图像质量的前提下,提高医学CT图像的可视化展示速度变得十分重要。由于医学断层图像数据间距较大,造成图像展示呈现阶梯状结构。插值计算是医学断层图像可视化过程中的一个重要步骤,然而,插值计算耗费大量时间,因此,高效的插值算法是提高断层图像可视化展示的一种关键。本文深入分析了光线投射法的基本原理。针对三线性插值方法在三维图像可视化绘制速度中的不足,将反距离加权插值方法应用到医学断层图像可视化中,通过对空间点集进行体素化、权值的计算以及插值点的映射,实现了反距离加权插值。实验结果表示,反距离加权算法更加灵活,算法运算次数少,在保证图像质量的前提下,降低了插值运算次数,提高了绘制速度。根据插值点间的空间关系,采用基于接近云算法的插值点间快速跳跃策略,有效地避免了大量的无效的插值计算。在接近云算法中,研究并设计了距离变换图,结合街区距离度量策略保证插值点沿着光线方向安全穿越空区域,跳到下一个插值点。实验表明,通过基于接近云算法的插值点间快速跳跃策略,降低了插值次数,提高了绘制速度。本课题开发了一个CT断层图像可视化展示平台,并将反距离加权插值与接近云算法应用于展示平台中,并取得了良好的可视化展示效果。更多还原

【Abstract】 Three-dimensional visualization for medical tomography image is an important application domain in visualization calculation. More and more medical imaging data are obtained from CT medical photograph device, and it is difficult to process so much data in three-dimensional visualization. Therefore, it is important to improve the speed of processing and ensure the image quality. Due to spacing among medical tomography images, ladderlike contour is displayed. Interpolation calculation is an important task in the visualization process for medical tomography image, and it costs a lot of time. Therefore, efficient interpolation algorithm is a key to improve speed for tomography image visualization.This paper analyzed the principle of Ray Casting. A new method of Inverse-Distance Weighting interpolation was proposed for medical tomography image visualization. It aimed at improving the speed of traditional Tri-linear interpolation in the three-dimensional image visualization. After Voxelization for point set in space, weight calculation and interpolation point mapping, Inverse-Distance Weighting was implemented. The experiment results showed that the proposed method was more flexible and less times of operation, then it reduced the number of interpolation and improved the imaging speed. In accordance with the spatial relationships among interpolation points, a fast skipping strategy based on Proximity Clouds algorithm was introduced, and it reduced a lot of invalid interpolation calculation. In the Proximity Clouds algorithm, we studied and designed the distance transformation map,which combines with city block distance metric strategyand eusures that the movement of interpolation point skipps the space area along the ray direction in order to reach the next point. The experiment proved that this method reduced the number of interpolation and improved the imaging speed. We developed visualization Platform for CT tomography image, and Inverse-Distance Weighting and Proximity Clouds algorithms were applied in the Platform, the running result showed that visualization was effective.更多还原