Robust Optical Flow Based Deformable Registration of Thoracic CT Images
Robust Optical Flow Based Deformable Registration of Thoracic CT Images
复制标题
基于鲁棒光流的胸部 CT 图像变形配准
DOI：
--
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
H. Bischof
中科院分区：
文献类型：
--
作者：
M. Urschler;Manuel Werlberger;E. Scheurer;H. Bischof
我们提出了一种基于数据和正则化项的鲁棒测量的光流变形配准方法。我们展示了该方法的两种具体实现，其中一种以各向同性方式对位移场中的梯度进行惩罚，另一种通过根据图像梯度各向异性地对惩罚进行加权来进行正则化。我们的数据项由标准光流约束的 L 范数组成。我们展示了一种数值算法，可以在原始对偶优化设置中求解两个提出的模型。我们的算法以多分辨率方式工作，并应用于 EMPIRE10 注册挑战的 20 个数据集。我们的结果显示了改进的空间。我们相当简单的模型不会惩罚非微分同胚变换，这会导致其中一项评估措施产生不良结果，并且它似乎不适合大变形情况。然而，我们的算法能够使用基于 CUDA 的专用 GPU 实现在几分钟内执行大约 400 个数据集大小的注册，与其他报告的算法相比，速度非常快。
We present an optical flow deformable registration method which is based on robust measures for data and regularization terms. We show two specific implementations of the method, where one penalizes gradients in the displacement field in an isotropic fashion and the other one regularizes by weighting the penalization according to the image gradients anisotropically. Our data term consists of the L-norm of the standard optical flow constraint. We show a numerical algorithm that solves the two proposed models in a primal-dual optimization setup. Our algorithm works in a multi-resolution manner and it is applied to the 20 data sets of the EMPIRE10 registration challenge. Our results show room for improvement. Our rather simple model does not penalize non-diffeomorphic transformations, which leads to bad results on one of the evaluation measures, and it seems unsuited for large deformations cases. However, our algorithm is able to perform registrations of data set sizes around 400 on the order of a few minutes using a dedicated CUDA based GPU implementation, which is very fast compared to other reported algorithms.