Dept. of Inf. Eng., Nagoya Univ.
IEEE Trans Med Imaging. 1990;9(3):270-80. doi: 10.1109/42.57764.
A system for craniofacial surgical planning utilizing stacks of 2-D tomographic images is described. The four parts of the system are image generation, 2-D surgical planning, 3-D plan confirmation, and rough prediction of face shape after an operation. The four parts are combined to provide a useful surgical planning system. Because a gradient shading technique is used for generating 3-D images, the bumpy shape of the voxel (volume element for 3-D objects) sometimes obscures the essential shape of the displayed objects. To smooth undesirable bumps without losing the essential object shape, 2-D filtering is applied. Arbitrary bone blocks can be specified and moved interactively with a graphic terminal to any derived location to aid surgical planning. The chosen plan can be confirmed by observing computer-generated 3-D images from arbitrary directions since the process necessary to generate a 3-D image for any one direction should be adequate for all directions. Postoperative face-shape prediction is available for evaluating the operation plan at the last stage of the planning sequence.
利用二维断层图像堆栈的颅面外科规划系统被描述。系统的四个部分包括图像生成、二维手术规划、三维计划确认和手术后面部形状的粗略预测。这四个部分结合在一起提供了一个有用的手术规划系统。由于使用了梯度阴影技术来生成三维图像，体素（三维对象的体积元素）的凹凸形状有时会掩盖显示对象的基本形状。为了在不丢失基本对象形状的情况下平滑不需要的凸起，可以应用二维滤波。可以指定任意的骨块并使用图形终端将其交互式移动到任何派生位置，以辅助手术规划。所选的计划可以通过从任意方向观察计算机生成的三维图像来确认，因为为任何一个方向生成三维图像所需的过程应该足以用于所有方向。在规划序列的最后阶段，可以进行术后面形预测，以评估手术计划。
