基于几何特征不变量的快速3D医学图像配准

针对颅部三维医学图像配准计算量大、配准效率低等问题,提出了一种基于几何特征空间约束的快速配准方法。提取三维轮廓点云,提出了一种基于点云集最优拟合环的特征构造方法,并以每个特征环和每个层的质心用作特征量,通过使用迭代最近点(Iterative Closest Point, ICP)方法完成快速配准。实验结果表明,与传统的ICP算法相比,该方法计算量小,配准精度高,配准速度快。它是一种有效的实时三维配准方法。

Fast 3D Medical Image Registration Based on Geometric Feature Invariants

Aiming at the of large amount of computational data and low registration efficiency in 3D cranial medical image registration, a fast registration method based on geometric feature space constraints is proposed. The algorithm extracts three-dimensional contour point clusters, and proposes a feature construction method based on the optimal fitting ring of point clusters. The feature rings and the centroids of each layer are used as feature quantities, and the fast registration is completed by using Iterative Closest Point (ICP) method. The experimental results show that the method has less computation amount, high satisfactory registration accuracy and much faster registration speed than the traditional ICP algorithm. It is an effective real-time three-dimensional registration method.