AI-VAD bbox post-processing

[djdameln]

Description

• This PR adds bounding box post-processing operations to the AI-VAD implementation. These operations were not mentioned in the paper, but were included in the original implementation.
• The post-processing operations consist of:
  • filtering by class label (some datasets benefit from only including person-regions)
  • filtering by bbox surface area
  • Removing overlapping bounding boxes
  • Adding regions that are undetected by the mask rcnn region extractor, but deemed to belong to the foreground by comparing pixel movement between consecutive video frames.
• Several configurable parameters were added to control the behaviour of the bbox post-processing
• All combined, these changes increase the image AUROC on ped2 dataset from 0.917 to 0.960

Changes

• Bug fix (non-breaking change which fixes an issue)
• Refactor (non-breaking change which refactors the code base)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

Checklist

• My code follows the pre-commit style and check guidelines of this project.
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing tests pass locally with my changes
• I have added a summary of my changes to the CHANGELOG (not for minor changes, docs and tests).

[samet-akcay]

Thanks, only a single comment.

[ashwinvaidya17]

Thanks! I have a few minor comments

[ashwinvaidya17]

I think the docstring needs to be updated

[ashwinvaidya17]

Thanks