Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Questions about code #5

Open
ZhileiChen99 opened this issue Oct 20, 2022 · 0 comments
Open

Questions about code #5

ZhileiChen99 opened this issue Oct 20, 2022 · 0 comments

Comments

@ZhileiChen99
Copy link

Thanks for your work. I read carefully about your source code and have some questions.

  1. neighborhood_consensus_loss: Why you use the relative coordinates of src_keypoints_knn and tgt_keypoints_knn to calculate the neighborhood_consensus_loss? This is different from the formulation in the paper.
  2. feature extractor: why the knn method does not exclude the center point? I think the exact number of neighbour is k-1.
  3. the input feature of DGCNN: Why the input feature is x-x_knn, rather than cat[x, x-x_knn]? This is different from the original DGCNN setting.
  4. I think the three questions have some relations, but I can't figure out it. I try to chang the knn method, but I find the network can't converge. For example, I exclude the center point in knn method, and I try to use cat[x, x-x_knn] as the input of DGCNN.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant