feat: add segmentation mask quality utilities#2263
Open
Aniketiitk21 wants to merge 2 commits into
Open
Conversation
|
|
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## develop #2263 +/- ##
========================================
Coverage 78% 78%
========================================
Files 66 67 +1
Lines 8374 8474 +100
========================================
+ Hits 6501 6597 +96
- Misses 1873 1877 +4 🚀 New features to boost your workflow:
|
Contributor
There was a problem hiding this comment.
Pull request overview
This PR adds single-mask segmentation quality primitives (region overlap + boundary-aware metrics) to complement existing batch utilities and CompactMask support, along with tests and documentation to make the new metrics discoverable and reliable.
Changes:
- Added
mask_iou,dice_coefficient,boundary_iou, andboundary_f_scoreutilities supporting 2D masks,(1, H, W)single-mask batches, andCompactMask(single mask). - Added unit tests covering overlap math, empty-mask semantics, tolerance behavior, dtype parity, and
CompactMaskparity. - Added a new docs page and MkDocs navigation entry for the mask metrics.
Assessment (n/5):
- Code quality: 4/5
- Testing: 3/5
- Docs: 4/5
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 5 comments.
Show a summary per file
| File | Description |
|---|---|
src/supervision/detection/utils/mask_metrics.py |
Implements new single-mask overlap/boundary metrics with validation and CompactMask support. |
tests/detection/utils/test_mask_metrics.py |
Adds unit tests for correctness, empty-mask semantics, tolerance behavior, dtype handling, and CompactMask parity. |
src/supervision/__init__.py |
Re-exports new metrics at the package top level and adds them to __all__. |
docs/detection/utils/mask_metrics.md |
Introduces documentation for the new metrics and their semantics. |
mkdocs.yml |
Adds the new Mask Metrics page to the documentation navigation. |
Comment on lines
+106
to
+109
| """ | ||
| Compute Intersection over Union (IoU) for a single pair of segmentation masks. | ||
|
|
||
| Args: |
Comment on lines
+149
to
+152
| """ | ||
| Compute Dice coefficient for a single pair of segmentation masks. | ||
|
|
||
| Args: |
Comment on lines
+196
to
+200
| Boundary IoU dilates the foreground contours of both masks by `tolerance` | ||
| pixels before measuring IoU. This makes the score more forgiving to small | ||
| contour shifts than standard region IoU. | ||
|
|
||
| Args: |
Comment on lines
+255
to
+259
| Boundary F-score measures contour agreement by matching predicted boundary | ||
| pixels to target boundary pixels within `tolerance` pixels, then combining | ||
| boundary precision and boundary recall into a single score. | ||
|
|
||
| Args: |
| assert math.isclose( | ||
| boundary_f_score(compact_prediction, compact_target, tolerance=1), | ||
| boundary_f_score(prediction, target, tolerance=1), | ||
| ) |
Borda
reviewed
May 26, 2026
| @@ -0,0 +1,59 @@ | |||
| --- | |||
| comments: true | |||
| status: new | |||
Member
There was a problem hiding this comment.
This needs to have decribtion as other pages
Borda
reviewed
May 26, 2026
| return None | ||
|
|
||
|
|
||
| def _extract_boundary(mask: npt.NDArray[np.bool_]) -> npt.NDArray[np.bool_]: |
Member
There was a problem hiding this comment.
also private funs shall have at leas summary docstring
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
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
Adds small, composable segmentation mask quality utilities for comparing a predicted binary mask with a target mask.
Type of Change
Motivation and Context
Supervision already provides
mask_iou_batchand strong segmentation support throughDetections.maskandCompactMask. This PR adds single-mask quality primitives that are easy to reuse in segmentation model debugging, annotation QA, dataset inspection, and contour-sensitive evaluation workflows.Changes Made
mask_iou,dice_coefficient,boundary_iou, andboundary_f_scoreCompactMaskparityTesting
Commands run:
python -m venv .venv ./.venv/Scripts/python.exe -m pip install -e . pytest ruff mypy ./.venv/Scripts/python.exe -m pip install mypy==1.15.0 ./.venv/Scripts/python.exe -m ruff check src/supervision/detection/utils/mask_metrics.py tests/detection/utils/test_mask_metrics.py src/supervision/__init__.py ./.venv/Scripts/python.exe -m mypy src/supervision/detection/utils/mask_metrics.py ./.venv/Scripts/python.exe -m pytest tests/detection/utils/test_mask_metrics.py -q ./.venv/Scripts/python.exe -m pytest src/supervision/detection/utils/mask_metrics.py -qGoogle Colab
Colab link: N/A
Screenshots/Videos
N/A
Additional Notes
Semantics are explicit and documented:
1.00.0toleranceis an integer pixel distance for boundary matchingThis PR is intentionally focused as PR 1 of a broader segmentation quality toolkit roadmap and does not include reports, error maps, or visualization helpers yet.