Skip to content

Convex Optimization method for point cloud registration

Notifications You must be signed in to change notification settings

tanghme0w/convopt4pcd

Repository files navigation

convopt4pcd

The functionality of each script should be clear by their name. Most of the scripts should be runnable as-is. Please feel free to contact us if you have any questions.

  • ransac, 3ds_ransac: implementation of RANSAC algorithm. ransac uses FPFH features, while 3ds_ransac uses 3DSmoothNet features.
  • kabsch: implementation of Kabsch algorithm.
  • sdrsac: implementation of CVPR 2019 SDRSAC paper.
  • Other scripts are for testing(e.g., show_result), format conversion(e.g., ply2mat, salient2mat, feat2mat), and exploration (e.g., test_rotation, render_salient) purposes.

necessary data files:

data
├─3ds-feature
│  ├─128dim
│  └─32dim
├─bunny-pcd
├─room-pcd
└─temple-pcd

About

Convex Optimization method for point cloud registration

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages