Add background remove, resizing, and additional pipelines (#277, #271, #267) #278
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
This was
linked to
issues
Apr 15, 2025
dlichy
force-pushed
the
issue_271_meshroom_pipelines
branch
from
e3167f7 to
18d8636
Compare
This commit also changes the signature of run_reconstruction to:
run_reconstruction(images: list[Path],
pipeline_name: str = "default_pipeline"
)-> Tuple[Photoscan, Path]
Here, pipeline_name is now a string representing the name of a .mg
pipeline in the meshroom_pipelines folder.
These warnings break the CI on windows11 python3.12, because we configure our pytest to treat all warnings as errors. There is a chance that the ONNX runtime support limitation warnings are not actually harmful to the user and just indicate a lack of official testing and support from Microsoft: microsoft/onnxruntime#20001 (comment)
ebrahimebrahim
force-pushed
the
issue_271_meshroom_pipelines
branch
from
ad0cb71 to
2ee2676
Compare
ebrahimebrahim
force-pushed
the
issue_271_meshroom_pipelines
branch
from
2ee2676 to
09b88a4
Compare
ebrahimebrahim
requested changes
Apr 17, 2025
Collaborator
ebrahimebrahim
left a comment
ebrahimebrahim
left a comment
There was a problem hiding this comment.
Amazing work! Things ran very smoothly. The code is clean. Several changes requested, including two sections for which I think unit testing will pay off.
Note that I force-pushed to rebase to main, and I also pushed a couple of minor things, so please reset your local version of this branch to here before you continue working. I won't mess with the branch again without prior warning 😄
| """Get the MODNet checkpoint path. Download it if not present. | ||
| """ | ||
| package = "openlifu.trk.modnet_checkpoints" | ||
| filename = "modnet_photographic_portrait_matting.onnx" |
Collaborator
There was a problem hiding this comment.
let's add src/openlifu/trk/modnet_checkpoints/modnet_photographic_portrait_matting.onnx to the .gitignore at the top level of this repo
or maybe more generally just *.onnx?
added 5 commits
April 17, 2025 14:37
ebrahimebrahim
force-pushed
the
issue_271_meshroom_pipelines
branch
from
cf1718e to
9e8786b
Compare
ebrahimebrahim
approved these changes
Apr 18, 2025
Collaborator
ebrahimebrahim
left a comment
ebrahimebrahim
left a comment
There was a problem hiding this comment.
everything works seamlessly, and many thanks for the unit tests!
Comment on lines
+172
to
+176
| # Test inverse is consistent | ||
| for orientation in range(9): | ||
| img_tform = apply_exif_orientation_numpy(img, orientation) | ||
| img_recon = apply_exif_orientation_numpy(img_tform, orientation, inverse=True) | ||
| np.testing.assert_array_equal(img, img_recon) |
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.
This pull request adds several new capabilities to the
run_reconstructionfunction:(Issue Downsample/Normalize resolution of photoscan reconstruction inputs #277) Adds a new argument
new_widththat resizes the images to the specified width while maintaining the aspect ratio.(Issue Photoscan background removal #267) Adds an argument
use_mask. When set toTrue, this runs MODNet on each image to obtain a background mask. The mask is then used to filter the depth maps during reconstruction with Meshroom.To use this feature:
MODNET_PATHto the path of the downloaded checkpoint, e.g.:(Issue Meshroom pipeline optimization #271) Adds a new argument
pipeline_namethat specifies one of the custom pipelines located in themeshroom_pipelinesfolder.