基于薄板样条和形状内容的医学图像非刚性配准方法研究

Non-rigid Medical Image Registration Based on the Thin-plate Spline and Shape Context

目的针对医学图像非刚性点配准的现状,给出一种基于点特征的非刚性配准方法。方法利用一种新的相似度测量方法——形状内容来解决两幅图像中点的对应关系,并利用点对应关系来估计非刚性映射函数。结果利用薄板样条实现了医学图像的快速准确非刚性配准。结论实验结果表明,上述方法获得了很好的配准效果。

Objective To present situation of medical image non-rigid registration, based on point-sets. Methods A new method to measure the similarity-the shape context was attached to solve the correspondences between points on the two shapes; and the correspondence was used to estimate the non-rigid transform function. Results The thin-plate spline was used to realize fast and exact non-rigid registration of medical image. Conclusion The experimental results show that this method can achieves better matching effect.