Non-rigid point set registration based on local neighborhood information support

作者：

Highlights：

• We proposed a novel non-rigid point set registration method that can preserve global and local structures.

• A new effective method to calculate the support function from neighbors is used and the more accurate initial correspondences can be obtained.

• The support contains two parts. The first part contains the matched neighbors of points. The second part contains the neighbors with the most similar relative spatial position.

• The performance of our method outperforms state-of-the-art methods in this area.

摘要

•We proposed a novel non-rigid point set registration method that can preserve global and local structures.•A new effective method to calculate the support function from neighbors is used and the more accurate initial correspondences can be obtained.•The support contains two parts. The first part contains the matched neighbors of points. The second part contains the neighbors with the most similar relative spatial position.•The performance of our method outperforms state-of-the-art methods in this area.

论文关键词：Non-rigid point set registration,Gaussian mixture model,Expectation–Maximization method,Local neighborhood information

论文评审过程：Received 19 August 2020, Revised 10 April 2022, Accepted 1 August 2022, Available online 2 August 2022, Version of Record 6 August 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108952