Mining consistent correspondences using co-occurrence statistics

作者：

Highlights：

• We integrate co occurrence statistics to construct more significant neighborhoods for each correspondence.

• We propose a novel non parametric mismatch removal algorithm based on co occurrence statistics.

• We propose a guided sampling method to significantly improve the quality of minimal subsets.

• Experimental results show the proposed method s are superior to some state of the art fitting methods.

摘要

•We integrate co occurrence statistics to construct more significant neighborhoods for each correspondence.•We propose a novel non parametric mismatch removal algorithm based on co occurrence statistics.•We propose a guided sampling method to significantly improve the quality of minimal subsets.•Experimental results show the proposed method s are superior to some state of the art fitting methods.

论文关键词：Feature matching,Geometric model fitting,Co-occurrence statistics,Guided sampling

论文评审过程：Received 25 September 2020, Revised 6 March 2021, Accepted 18 May 2021, Available online 27 May 2021, Version of Record 4 June 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108062