Difference between new_points and grouped_points
popopochan opened this issue · 5 comments
@ma-xu
I'm sorry to keep coming back to you.
Local_grouper calculates and concats new_xyz and new_points, and grouped_xyz and grouped_points, but I would like to know the difference between these two.
Also, what is the difference between fps_idx and idx?
Thank you.
fps_idx denotes the idx of sampling center points, idx denote the idx of neighbors of center points
@cccchensir Thanks for your explanation.
@popopochan please see ccchensir's explanation. That's right.
Dear @cccchensir @ma-xu
Thank you for responses.
I have one more question.
In the geometric affine module, α and β are learnable parameters, and in the first step, α is all 1 and β is all 0, but I would appreciate it if you could tell me the process, reason, and advantage of changing it at each step.
Also, as a result, is the output point cloud affine transformed from the input point cloud?
I am a beginner. Please.
Thank you.
they are leanrable parameters implemented by nn.Parameter, will be automatically optimized and updated like parameters in learnable layers.
Closed. Feel free to reopen it if you have any further questions or concerns.