depth grad, and add test #126
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
wanmeihuali
reviewed
Sep 18, 2023
taichi_3d_gaussian_splatting/GaussianPointCloudRasterisation.py
Outdated
Show resolved
Hide resolved
taichi_3d_gaussian_splatting/GaussianPointCloudRasterisation.py
Outdated
Show resolved
Hide resolved
Comment on lines
629
to
632
| alpha_grad_from_depth = (T_i * (depth_i-d_pixel) + 1.0/(1.0-alpha) * \ | ||
| (depth_w_i - acc_alpha_w_i* d_pixel)) / \ | ||
| (org_accumulated_alpha+0.00001) * pixel_depth_grad | ||
| alpha_grad_from_accumulated_alpha = (T_i- 1.0/(1.0-alpha) * acc_alpha_w_i) * _accumulated_alpha_grad |
There was a problem hiding this comment.
Please format your code. I haven't looked deep into your formula so I'm not sure if it is correct, but for implementation, it seems to be the same as your formula. Have you gotten it to converge on alpha in some more complex tests? e.g. as an extra loss?
There was a problem hiding this comment.
Formated the code, hopefully it's now easier to understand.
Have you gotten it to converge on alpha in some more complex tests? e.g. as an extra loss?
What do you mean as an "extra loss"? Do you mean train with RGB and depth loss together?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Test depth grad on xyz and alpha.
Note: I found the initialization of alpha is important:
Results: Optimize pts alpha only (red are all points, blue are points with high alpha,)
Results: Optimize pts coords only
