【24h】

Fast non-rigid image registration using viscous fluid model and B-spline

机译:使用粘性流体模型和B样条快速进行非刚性图像配准

获取原文
获取原文并翻译 | 示例

摘要

Viscous fluid based non-rigid registration algorithm is an appropriate method for registering objects with large-scale difference. The core component of the algorithm is to solve the fluid partial differential equations. Usually, iterative methods, such as successive over-relaxation, are employed to obtain the solution, but the computational complexity is very tremendous. To reduce the complexity, a fast method based on fluid model combined with Bspline is presented. Firstly, B-spline is used to model the velocity fields, so that the large number of unknowns are reduced to fewer B-spline coefficients; Secondly, by use of some special attributes of B-spline and fast Fourier transform(FFT), direct inversion for the B-spline coefficients are deduced. Experimental results show that our new method is not only suitable for the large-scale image deformation but also amounts relatively lower computational costs.
机译:基于粘性流体的非刚性配准算法是一种用于配准具有较大差异的对象的合适方法。该算法的核心部分是求解流体偏微分方程。通常,采用迭代方法(例如连续过度松弛)来获得解决方案,但是计算复杂性非常高。为了降低复杂度,提出了一种基于流体模型结合Bspline的快速方法。首先,使用B样条对速度场进行建模,从而将大量未知数减少为更少的B样条系数。其次,利用B样条的一些特殊属性和快速傅里叶变换(FFT),推导了B样条系数的直接反演。实验结果表明,该方法不仅适用于大规模图像变形,而且计算量相对较低。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号