一种改进的基于Facet模型的亚体素表面检测算法

在基于CT的工业应用中，常常需要从CT切片序列获得实体的3D高精度表面信息。介绍了一种基于Facet模型的亚体素表面检测算法，并针对其计算效率低的问题，给出了一种3DFacet模型的加速算法。该改进算法将Facet模型的三维卷积核分解为3个一维卷积核，使算法复杂度从O（m^3）降低到O（3m），并采用增量算法解决了由此产生的存储空间问题。通过对仿真图像的实验，验证了该算法的效率和精度，表明该算法在保持原始算法精度的同时，处理速度提高了约两倍。

An Improved Subvoxel Surface Detection Algorithm Based on Facet Model

In industrial applications based on CT, three-dimensional surface information in high accuracy of an entity needs to be extracted from CT slice images. A subvoxel surface detection algorithm based on Facet model was introduced. For the low efficiency of the algorithm, an accelerated method was provided for computing 3D Facet model. The improved one reduces the computation complexity from O(m~3) to O(3m) by decomposing the 3D convolution mask to three 1D masks, and adopts an incremental strategy to solve the subsequent high memory consuming problem. Finally, experiments on simulated images validate the efficiency and accuracy of the algorithm. The results show that the improved method can increase the speed about 2 times, with accuracy remaining the same.