Mesh R-CNN Mesh Losses #16
Merged
Conversation
This file contains hidden or 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
Created a script that is very similar to the vizualize_mesh but also includes support for displaying the normal vectors. Images displaying the normal vectors are also provided in this commit for demonstration of progress but will be deleted later.
Create voxels will also be used for testing the mesh sampler.
This was done so the mesh sampler and the utility functions/class for getting the closest points in each point cloud can both live in mesh sample while providing a clear distinction between them.
Note: this is currently a prototype. A rationale for pushing this prototyping and the TODOs are listed in the module docstring.
Added class that encompases all of the mesh based loss terms (chamfer, normals, and edge) as well as voxel BCE loss. Added option to pass batch weights to chamfer and normal loss. Refactored some code for greater modularity.
TF implementation was applying point reduction before multiplying by weights when the original implementation was doing the opposite.
These tests now run faster and no longer assert values for functions that depend on other random functions (i.e. the mesh sampler).
davidliii
reviewed
Mar 28, 2022
Collaborator
davidliii
left a comment
There was a problem hiding this comment.
Please add a config dataclass for the loss parameters (example YoloLoss)
Wenxin-Jiang
pushed a commit
that referenced
this pull request
Jan 12, 2023
Use `depthwise_separable_conv2d`
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Outdated suggestions cannot be applied.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.
Description
Implements the necessary components to compute the voxel and mesh losses.
The primary additions includes the:
Type of change
Tests
The tests added can be found in:
mesh_losses_differential_test.py: Unit tests for the mesh sampler.mesh_losses_test: Unit tests for the mesh losses.mesh_losses_differential_test: Differential tests for the mesh losses.Checklist