Hyungtae Lim

If there are no further questions, we will close the issue :) Thx

Voxelized-FPFH works well even though we do not extract any keypoints. Please follow the instructions in the papers for 3D sparse point clouds, not the Bunny dataset: - https://arxiv.org/abs/2203.06612 -...

If there are no further questions, we will close the issue :) Thx

Yes, I think 4K of correspondences are too many (empirically, TEASER++ works in near-real time under 600 correspondences). Please filter the correspondences before you give them as inputs.

If there are no further questions, we will close the issue :) Thx

Yes, 180K points are too many. Empirically, TEASER++ can cover up to < 700-800 points

If there are no further questions, we will close the issue :) Thx

What do you mean "ideal"? Do you mean the error is not equal to zero? It's the natural characteristic of a non-minimal solver. Because the problem we want to solve...

In short, it does not go to global optimum to reduce the effect of outliers (because owing to the outliers, the actual global optimum != global optimum in the loss...

Please run `sudo ldconfig` after the `sudo make install`!