Add anchor pruning to projects documentation + Discussion about supporting the method in open-mmlab toolbox. #8120
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Outdated suggestions cannot be applied.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.
Dear MMDetection maintainers,
I have recently published a journal article about "Anchor Pruning for Object Detection." (https://doi.org/10.1016/j.cviu.2022.103445). As it was built using MMDetection I would like to contribute back to the main repository.
This pull requests consists of 2 parts:
Part 1 is the request to be added to the project list as can be seen in the changes of this pull request.
Part 2 is a discussion about whether you're open to supporting anchor pruning in the open-mmlab toolbox.
This would require 2 changes: 1) my adaption makes the AnchorGenerator more general to freely define each anchor in anchor-based object detectors. This would be a change that is fully compatible with all current models released by MMDetection and it would also open up the opportunity to match the SSD-Lite anchors precisely to the ones in the official TF repository.
2) The actual anchor pruning algorithm does seem to be a good candidate to be merged into the MMRazor toolbox. This would require a rewrite of my algorithm to match the templates of this repository but if you're open to supporting my pruning method than I'm happy to adapt my code.
Kind regards,
Maxim Bonnaerens