保拓扑结构的非刚性医学图像配准方法

Topology-Preserving Non-Rigid Medical Image Registration Method

摘要: 图像配准在图像处理领域发挥着举足轻重的作用，尤其是在医学图像处理领域，需要能保持原器官拓扑结构的非刚性图像配准方法。本文基于类Beltrami系数的保拓扑结构的非刚性医学图像配准方法，提出了新的惩罚函数，使得类Beltrami系数满足$N\left(y\right)<1$ ，能更好地保证配准过程中的拓扑结构的保持。同时，我们提出了新的保拓扑结构的非刚性医学图像配准模型，并证明了关键数学性质。最后，我们在多个复杂的实际场景中进行了大量实验，验证了我们方法的优越性。

Abstract: Image registration plays a crucial role in the field of image processing. Particularly in the area of medical image processing, there is a need for non-rigid image registration methods that can preserve the topological structure of original organs. This paper presents a novel penalty function based on the topological-structure-preserving non-rigid medical image registration method using the Beltrami-like coefficients. By using this penalty function, the Beltrami-like coefficients satisfy the condition $N\left(y\right)<1$ , which can better ensure the preservation of the topological structure during the registration process. Meanwhile, we propose a new topological-structure-preserving non-rigid medical image registration model and prove its key mathematical properties. Finally, we conduct numerous experiments in multiple complex real-world scenarios to verify the superiority of our method.