diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 9465c2af..fb85a294 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -45,7 +45,7 @@ repos: - repo: https://github.com/astral-sh/ruff-pre-commit - rev: v0.4.4 + rev: v0.4.8 hooks: - id: ruff args: [--fix, --exit-non-zero-on-fix] diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 0003f1aa..d32ab0cf 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -64,7 +64,7 @@ To run the pre-commit tool, follow these steps: 3. Run the command `pre-commit run --all-files`. This will execute the pre-commit hooks configured for this project against the modified files. If any issues are found, the pre-commit tool will provide feedback on how to resolve them. Make the necessary changes and re-run the pre-commit command until all issues are resolved. -4. You can also install pre-commit as a git hook by execute `pre-commit install`. Every time you made `git commit` pre-commit run automatically for you. +4. You can also install pre-commit as a git hook by executing `pre-commit install`. Every time you do a `git commit` pre-commit run automatically for you. ### Docstrings diff --git a/docs/changelog.md b/docs/changelog.md index 47d160aa..024edd1a 100644 --- a/docs/changelog.md +++ b/docs/changelog.md @@ -1,3 +1,83 @@ +### 0.21.0 Jun 5, 2024 + +- Added [#500](https://github.com/roboflow/supervision/pull/500): [`sv.Detections.with_nmm`](https://supervision.roboflow.com/develop/detection/core/#supervision.detection.core.Detections.with_nmm) to perform non-maximum merging on the current set of object detections. + +- Added [#1221](https://github.com/roboflow/supervision/pull/1221): [`sv.Detections.from_lmm`](https://supervision.roboflow.com/develop/detection/core/#supervision.detection.core.Detections.from_lmm) allowing to parse Large Multimodal Model (LMM) text result into [`sv.Detections`](https://supervision.roboflow.com/develop/detection/core/) object. For now `from_lmm` supports only [PaliGemma](https://colab.research.google.com/github/roboflow-ai/notebooks/blob/main/notebooks/how-to-finetune-paligemma-on-detection-dataset.ipynb) result parsing. + +```python +import supervision as sv + +paligemma_result = " cat" +detections = sv.Detections.from_lmm( + sv.LMM.PALIGEMMA, + paligemma_result, + resolution_wh=(1000, 1000), + classes=['cat', 'dog'] +) +detections.xyxy +# array([[250., 250., 750., 750.]]) + +detections.class_id +# array([0]) +``` + +- Added [#1236](https://github.com/roboflow/supervision/pull/1236): [`sv.VertexLabelAnnotator`](https://supervision.roboflow.com/develop/keypoint/annotators/#supervision.keypoint.annotators.EdgeAnnotator.annotate) allowing to annotate every vertex of a keypoint skeleton with custom text and color. + +```python +import supervision as sv + +image = ... +key_points = sv.KeyPoints(...) + +edge_annotator = sv.EdgeAnnotator( + color=sv.Color.GREEN, + thickness=5 +) +annotated_frame = edge_annotator.annotate( + scene=image.copy(), + key_points=key_points +) +``` + +- Added [#1147](https://github.com/roboflow/supervision/pull/1147): [`sv.KeyPoints.from_inference`](https://supervision.roboflow.com/develop/keypoint/core/#supervision.keypoint.core.KeyPoints.from_inference) allowing to create [`sv.KeyPoints`](https://supervision.roboflow.com/develop/keypoint/core/#supervision.keypoint.core.KeyPoints) from [Inference](https://github.com/roboflow/inference) result. + +- Added [#1138](https://github.com/roboflow/supervision/pull/1138): [`sv.KeyPoints.from_yolo_nas`](https://supervision.roboflow.com/develop/keypoint/core/#supervision.keypoint.core.KeyPoints.from_yolo_nas) allowing to create [`sv.KeyPoints`](https://supervision.roboflow.com/develop/keypoint/core/#supervision.keypoint.core.KeyPoints) from [YOLO-NAS](https://github.com/Deci-AI/super-gradients/blob/master/YOLONAS.md) result. + +- Added [#1163](https://github.com/roboflow/supervision/pull/1163): [`sv.mask_to_rle`](https://supervision.roboflow.com/develop/datasets/utils/#supervision.dataset.utils.rle_to_mask) and [`sv.rle_to_mask`](https://supervision.roboflow.com/develop/datasets/utils/#supervision.dataset.utils.rle_to_mask) allowing for easy conversion between mask and rle formats. + +- Changed [#1236](https://github.com/roboflow/supervision/pull/1236): [`sv.InferenceSlicer`](https://supervision.roboflow.com/develop/detection/tools/inference_slicer/) allowing to select overlap filtering strategy (`NONE`, `NON_MAX_SUPPRESSION` and `NON_MAX_MERGE`). + +- Changed [#1178](https://github.com/roboflow/supervision/pull/1178): [`sv.InferenceSlicer`](https://supervision.roboflow.com/develop/detection/tools/inference_slicer/) adding instance segmentation model support. + +```python +import cv2 +import numpy as np +import supervision as sv +from inference import get_model + +model = get_model(model_id="yolov8x-seg-640") +image = cv2.imread() + +def callback(image_slice: np.ndarray) -> sv.Detections: + results = model.infer(image_slice)[0] + return sv.Detections.from_inference(results) + +slicer = sv.InferenceSlicer(callback = callback) +detections = slicer(image) + +mask_annotator = sv.MaskAnnotator() +label_annotator = sv.LabelAnnotator() + +annotated_image = mask_annotator.annotate( + scene=image, detections=detections) +annotated_image = label_annotator.annotate( + scene=annotated_image, detections=detections) +``` + +- Changed [#1228](https://github.com/roboflow/supervision/pull/1228): [`sv.LineZone`](https://supervision.roboflow.com/develop/detection/tools/line_zone/) making it 10-20 times faster, depending on the use case. + +- Changed [#1163](https://github.com/roboflow/supervision/pull/1163): [`sv.DetectionDataset.from_coco`](https://supervision.roboflow.com/develop/datasets/core/#supervision.dataset.core.DetectionDataset.from_coco) and [`sv.DetectionDataset.as_coco`](https://supervision.roboflow.com/develop/datasets/core/#supervision.dataset.core.DetectionDataset.as_coco) adding support for run-length encoding (RLE) mask format. + ### 0.20.0 April 24, 2024 - Added [#1128](https://github.com/roboflow/supervision/pull/1128): [`sv.KeyPoints`](/0.20.0/keypoint/core/#supervision.keypoint.core.KeyPoints) to provide initial support for pose estimation and broader keypoint detection models. diff --git a/docs/cookbooks.md b/docs/cookbooks.md index dd963edb..6f87958f 100644 --- a/docs/cookbooks.md +++ b/docs/cookbooks.md @@ -1,7 +1,6 @@ --- template: cookbooks.html comments: true -status: new hide: - navigation - toc diff --git a/docs/detection/annotators.md b/docs/detection/annotators.md index 958f2a74..4d912cae 100644 --- a/docs/detection/annotators.md +++ b/docs/detection/annotators.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # Annotators diff --git a/docs/detection/double_detection_filter.md b/docs/detection/double_detection_filter.md new file mode 100644 index 00000000..1631852f --- /dev/null +++ b/docs/detection/double_detection_filter.md @@ -0,0 +1,30 @@ +--- +comments: true +status: new +--- + +# Double Detection Filter + +
+

OverlapFilter

+
+ +:::supervision.detection.overlap_filter.OverlapFilter + +
+

box_non_max_suppression

+
+ +:::supervision.detection.overlap_filter.box_non_max_suppression + +
+

mask_non_max_suppression

+
+ +:::supervision.detection.overlap_filter.mask_non_max_suppression + +
+

box_non_max_merge

+
+ +:::supervision.detection.overlap_filter.box_non_max_merge diff --git a/docs/detection/tools/inference_slicer.md b/docs/detection/tools/inference_slicer.md index 5d5d08bc..7a5d3e57 100644 --- a/docs/detection/tools/inference_slicer.md +++ b/docs/detection/tools/inference_slicer.md @@ -1,5 +1,6 @@ --- comments: true +status: new --- # InferenceSlicer diff --git a/docs/detection/tools/line_zone.md b/docs/detection/tools/line_zone.md index 8d13822c..22e9c0a8 100644 --- a/docs/detection/tools/line_zone.md +++ b/docs/detection/tools/line_zone.md @@ -1,5 +1,6 @@ --- comments: true +status: new ---
diff --git a/docs/detection/tools/save_detections.md b/docs/detection/tools/save_detections.md index a82ce5df..a24cee57 100644 --- a/docs/detection/tools/save_detections.md +++ b/docs/detection/tools/save_detections.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # Save Detections diff --git a/docs/detection/utils.md b/docs/detection/utils.md index f9c9473b..ea98c868 100644 --- a/docs/detection/utils.md +++ b/docs/detection/utils.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # Detection Utils @@ -17,18 +16,6 @@ status: new :::supervision.detection.utils.mask_iou_batch -
-

box_non_max_suppression

-
- -:::supervision.detection.utils.box_non_max_suppression - -
-

mask_non_max_suppression

-
- -:::supervision.detection.utils.mask_non_max_suppression -

polygon_to_mask

diff --git a/docs/how_to/detect_and_annotate.md b/docs/how_to/detect_and_annotate.md index a9a4405e..52e3174b 100644 --- a/docs/how_to/detect_and_annotate.md +++ b/docs/how_to/detect_and_annotate.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # Detect and Annotate diff --git a/docs/how_to/save_detections.md b/docs/how_to/save_detections.md index 94de6c61..05d5faad 100644 --- a/docs/how_to/save_detections.md +++ b/docs/how_to/save_detections.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # Save Detections diff --git a/docs/trackers.md b/docs/trackers.md index 47f70061..cb44441f 100644 --- a/docs/trackers.md +++ b/docs/trackers.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # ByteTrack diff --git a/docs/utils/image.md b/docs/utils/image.md index 8f170d35..8e39136a 100644 --- a/docs/utils/image.md +++ b/docs/utils/image.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # Image Utils @@ -12,7 +11,7 @@ status: new :::supervision.utils.image.crop_image
-

crop_image

+

scale_image

:::supervision.utils.image.scale_image diff --git a/docs/utils/iterables.md b/docs/utils/iterables.md index b65cd954..5ae92dc9 100644 --- a/docs/utils/iterables.md +++ b/docs/utils/iterables.md @@ -1,6 +1,5 @@ --- comments: true -status: new --- # Iterables Utils diff --git a/mkdocs.yml b/mkdocs.yml index f257238d..96728971 100644 --- a/mkdocs.yml +++ b/mkdocs.yml @@ -35,19 +35,20 @@ extra_css: nav: - - Home: index.md - - How to: + - Supervision: index.md + - Learn: - Detect and Annotate: how_to/detect_and_annotate.md - Save Detections: how_to/save_detections.md - Filter Detections: how_to/filter_detections.md - Detect Small Objects: how_to/detect_small_objects.md - Track Objects on Video: how_to/track_objects.md - - API: + - Reference - Code API: - Detection and Segmentation: - Core: detection/core.md - Annotators: detection/annotators.md - Metrics: detection/metrics.md + - Double Detection Filter: detection/double_detection_filter.md - Utils: detection/utils.md - Keypoint Detection: - Core: keypoint/core.md @@ -78,7 +79,7 @@ nav: - Contributing: contributing.md - Code of Conduct: code_of_conduct.md - License: license.md - - Changelog: + - Release Notes: - Changelog: changelog.md - Deprecated: deprecated.md diff --git a/poetry.lock b/poetry.lock index 0a9276f8..ecad075d 100644 --- a/poetry.lock +++ b/poetry.lock @@ -1253,21 +1253,21 @@ test-extra = ["curio", "matplotlib (!=3.2.0)", "nbformat", "numpy (>=1.21)", "pa [[package]] name = "ipywidgets" -version = "8.1.2" +version = "8.1.3" description = "Jupyter interactive widgets" optional = false python-versions = ">=3.7" files = [ - {file = "ipywidgets-8.1.2-py3-none-any.whl", hash = "sha256:bbe43850d79fb5e906b14801d6c01402857996864d1e5b6fa62dd2ee35559f60"}, - {file = "ipywidgets-8.1.2.tar.gz", hash = "sha256:d0b9b41e49bae926a866e613a39b0f0097745d2b9f1f3dd406641b4a57ec42c9"}, + {file = "ipywidgets-8.1.3-py3-none-any.whl", hash = "sha256:efafd18f7a142248f7cb0ba890a68b96abd4d6e88ddbda483c9130d12667eaf2"}, + {file = "ipywidgets-8.1.3.tar.gz", hash = "sha256:f5f9eeaae082b1823ce9eac2575272952f40d748893972956dc09700a6392d9c"}, ] [package.dependencies] comm = ">=0.1.3" ipython = ">=6.1.0" -jupyterlab-widgets = ">=3.0.10,<3.1.0" +jupyterlab-widgets = ">=3.0.11,<3.1.0" traitlets = ">=4.3.1" -widgetsnbextension = ">=4.0.10,<4.1.0" +widgetsnbextension = ">=4.0.11,<4.1.0" [package.extras] test = ["ipykernel", "jsonschema", "pytest (>=3.6.0)", "pytest-cov", "pytz"] @@ -1638,13 +1638,13 @@ test = ["hatch", "ipykernel", "openapi-core (>=0.18.0,<0.19.0)", "openapi-spec-v [[package]] name = "jupyterlab-widgets" -version = "3.0.10" +version = "3.0.11" description = "Jupyter interactive widgets for JupyterLab" optional = false python-versions = ">=3.7" files = [ - {file = "jupyterlab_widgets-3.0.10-py3-none-any.whl", hash = "sha256:dd61f3ae7a5a7f80299e14585ce6cf3d6925a96c9103c978eda293197730cb64"}, - {file = "jupyterlab_widgets-3.0.10.tar.gz", hash = "sha256:04f2ac04976727e4f9d0fa91cdc2f1ab860f965e504c29dbd6a65c882c9d04c0"}, + {file = "jupyterlab_widgets-3.0.11-py3-none-any.whl", hash = "sha256:78287fd86d20744ace330a61625024cf5521e1c012a352ddc0a3cdc2348becd0"}, + {file = "jupyterlab_widgets-3.0.11.tar.gz", hash = "sha256:dd5ac679593c969af29c9bed054c24f26842baa51352114736756bc035deee27"}, ] [[package]] @@ -2164,13 +2164,13 @@ requests = "*" [[package]] name = "mkdocs-git-revision-date-localized-plugin" -version = "1.2.5" +version = "1.2.6" description = "Mkdocs plugin that enables displaying the localized date of the last git modification of a markdown file." optional = false python-versions = ">=3.8" files = [ - {file = "mkdocs_git_revision_date_localized_plugin-1.2.5-py3-none-any.whl", hash = "sha256:d796a18b07cfcdb154c133e3ec099d2bb5f38389e4fd54d3eb516a8a736815b8"}, - {file = "mkdocs_git_revision_date_localized_plugin-1.2.5.tar.gz", hash = "sha256:0c439816d9d0dba48e027d9d074b2b9f1d7cd179f74ba46b51e4da7bb3dc4b9b"}, + {file = "mkdocs_git_revision_date_localized_plugin-1.2.6-py3-none-any.whl", hash = "sha256:f015cb0f3894a39b33447b18e270ae391c4e25275cac5a626e80b243784e2692"}, + {file = "mkdocs_git_revision_date_localized_plugin-1.2.6.tar.gz", hash = "sha256:e432942ce4ee8aa9b9f4493e993dee9d2cc08b3ea2b40a3d6b03ca0f2a4bcaa2"}, ] [package.dependencies] @@ -2199,13 +2199,13 @@ pygments = ">2.12.0" [[package]] name = "mkdocs-material" -version = "9.5.24" +version = "9.5.26" description = "Documentation that simply works" optional = false python-versions = ">=3.8" files = [ - {file = "mkdocs_material-9.5.24-py3-none-any.whl", hash = "sha256:e12cd75954c535b61e716f359cf2a5056bf4514889d17161fdebd5df4b0153c6"}, - {file = "mkdocs_material-9.5.24.tar.gz", hash = "sha256:02d5aaba0ee755e707c3ef6e748f9acb7b3011187c0ea766db31af8905078a34"}, + {file = "mkdocs_material-9.5.26-py3-none-any.whl", hash = "sha256:5d01fb0aa1c7946a1e3ae8689aa2b11a030621ecb54894e35aabb74c21016312"}, + {file = "mkdocs_material-9.5.26.tar.gz", hash = "sha256:56aeb91d94cffa43b6296fa4fbf0eb7c840136e563eecfd12c2d9e92e50ba326"}, ] [package.dependencies] @@ -2485,13 +2485,13 @@ setuptools = "*" [[package]] name = "notebook" -version = "7.2.0" +version = "7.2.1" description = "Jupyter Notebook - A web-based notebook environment for interactive computing" optional = false python-versions = ">=3.8" files = [ - {file = "notebook-7.2.0-py3-none-any.whl", hash = "sha256:b4752d7407d6c8872fc505df0f00d3cae46e8efb033b822adacbaa3f1f3ce8f5"}, - {file = "notebook-7.2.0.tar.gz", hash = "sha256:34a2ba4b08ad5d19ec930db7484fb79746a1784be9e1a5f8218f9af8656a141f"}, + {file = "notebook-7.2.1-py3-none-any.whl", hash = "sha256:f45489a3995746f2195a137e0773e2130960b51c9ac3ce257dbc2705aab3a6ca"}, + {file = "notebook-7.2.1.tar.gz", hash = "sha256:4287b6da59740b32173d01d641f763d292f49c30e7a51b89c46ba8473126341e"}, ] [package.dependencies] @@ -3039,13 +3039,13 @@ tomli = {version = ">=1.1.0", markers = "python_version < \"3.11\""} [[package]] name = "pytest" -version = "8.2.1" +version = "8.2.2" description = "pytest: simple powerful testing with Python" optional = false python-versions = ">=3.8" files = [ - {file = "pytest-8.2.1-py3-none-any.whl", hash = "sha256:faccc5d332b8c3719f40283d0d44aa5cf101cec36f88cde9ed8f2bc0538612b1"}, - {file = "pytest-8.2.1.tar.gz", hash = "sha256:5046e5b46d8e4cac199c373041f26be56fdb81eb4e67dc11d4e10811fc3408fd"}, + {file = "pytest-8.2.2-py3-none-any.whl", hash = "sha256:c434598117762e2bd304e526244f67bf66bbd7b5d6cf22138be51ff661980343"}, + {file = "pytest-8.2.2.tar.gz", hash = "sha256:de4bb8104e201939ccdc688b27a89a7be2079b22e2bd2b07f806b6ba71117977"}, ] [package.dependencies] @@ -3461,13 +3461,13 @@ files = [ [[package]] name = "requests" -version = "2.32.2" +version = "2.32.3" description = "Python HTTP for Humans." optional = false python-versions = ">=3.8" files = [ - {file = "requests-2.32.2-py3-none-any.whl", hash = "sha256:fc06670dd0ed212426dfeb94fc1b983d917c4f9847c863f313c9dfaaffb7c23c"}, - {file = "requests-2.32.2.tar.gz", hash = "sha256:dd951ff5ecf3e3b3aa26b40703ba77495dab41da839ae72ef3c8e5d8e2433289"}, + {file = "requests-2.32.3-py3-none-any.whl", hash = "sha256:70761cfe03c773ceb22aa2f671b4757976145175cdfca038c02654d061d6dcc6"}, + {file = "requests-2.32.3.tar.gz", hash = "sha256:55365417734eb18255590a9ff9eb97e9e1da868d4ccd6402399eaf68af20a760"}, ] [package.dependencies] @@ -3662,28 +3662,28 @@ files = [ [[package]] name = "ruff" -version = "0.4.5" +version = "0.4.8" description = "An extremely fast Python linter and code formatter, written in Rust." optional = false python-versions = ">=3.7" files = [ - {file = "ruff-0.4.5-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:8f58e615dec58b1a6b291769b559e12fdffb53cc4187160a2fc83250eaf54e96"}, - {file = "ruff-0.4.5-py3-none-macosx_11_0_arm64.whl", hash = "sha256:84dd157474e16e3a82745d2afa1016c17d27cb5d52b12e3d45d418bcc6d49264"}, - {file = "ruff-0.4.5-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:25f483ad9d50b00e7fd577f6d0305aa18494c6af139bce7319c68a17180087f4"}, - {file = "ruff-0.4.5-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:63fde3bf6f3ad4e990357af1d30e8ba2730860a954ea9282c95fc0846f5f64af"}, - {file = "ruff-0.4.5-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:78e3ba4620dee27f76bbcad97067766026c918ba0f2d035c2fc25cbdd04d9c97"}, - {file = "ruff-0.4.5-py3-none-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:441dab55c568e38d02bbda68a926a3d0b54f5510095c9de7f95e47a39e0168aa"}, - {file = "ruff-0.4.5-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:1169e47e9c4136c997f08f9857ae889d614c5035d87d38fda9b44b4338909cdf"}, - {file = "ruff-0.4.5-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:755ac9ac2598a941512fc36a9070a13c88d72ff874a9781493eb237ab02d75df"}, - {file = "ruff-0.4.5-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f4b02a65985be2b34b170025a8b92449088ce61e33e69956ce4d316c0fe7cce0"}, - {file = "ruff-0.4.5-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:75a426506a183d9201e7e5664de3f6b414ad3850d7625764106f7b6d0486f0a1"}, - {file = "ruff-0.4.5-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:6e1b139b45e2911419044237d90b60e472f57285950e1492c757dfc88259bb06"}, - {file = "ruff-0.4.5-py3-none-musllinux_1_2_i686.whl", hash = "sha256:a6f29a8221d2e3d85ff0c7b4371c0e37b39c87732c969b4d90f3dad2e721c5b1"}, - {file = "ruff-0.4.5-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:d6ef817124d72b54cc923f3444828ba24fa45c3164bc9e8f1813db2f3d3a8a11"}, - {file = "ruff-0.4.5-py3-none-win32.whl", hash = "sha256:aed8166c18b1a169a5d3ec28a49b43340949e400665555b51ee06f22813ef062"}, - {file = "ruff-0.4.5-py3-none-win_amd64.whl", hash = "sha256:b0b03c619d2b4350b4a27e34fd2ac64d0dabe1afbf43de57d0f9d8a05ecffa45"}, - {file = "ruff-0.4.5-py3-none-win_arm64.whl", hash = "sha256:9d15de3425f53161b3f5a5658d4522e4eee5ea002bf2ac7aa380743dd9ad5fba"}, - {file = "ruff-0.4.5.tar.gz", hash = "sha256:286eabd47e7d4d521d199cab84deca135557e6d1e0f0d01c29e757c3cb151b54"}, + {file = "ruff-0.4.8-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:7663a6d78f6adb0eab270fa9cf1ff2d28618ca3a652b60f2a234d92b9ec89066"}, + {file = "ruff-0.4.8-py3-none-macosx_11_0_arm64.whl", hash = "sha256:eeceb78da8afb6de0ddada93112869852d04f1cd0f6b80fe464fd4e35c330913"}, + {file = "ruff-0.4.8-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:aad360893e92486662ef3be0a339c5ca3c1b109e0134fcd37d534d4be9fb8de3"}, + {file = "ruff-0.4.8-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:284c2e3f3396fb05f5f803c9fffb53ebbe09a3ebe7dda2929ed8d73ded736deb"}, + {file = "ruff-0.4.8-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a7354f921e3fbe04d2a62d46707e569f9315e1a613307f7311a935743c51a764"}, + {file = "ruff-0.4.8-py3-none-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:72584676164e15a68a15778fd1b17c28a519e7a0622161eb2debdcdabdc71883"}, + {file = "ruff-0.4.8-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:9678d5c9b43315f323af2233a04d747409d1e3aa6789620083a82d1066a35199"}, + {file = "ruff-0.4.8-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:704977a658131651a22b5ebeb28b717ef42ac6ee3b11e91dc87b633b5d83142b"}, + {file = "ruff-0.4.8-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d05f8d6f0c3cce5026cecd83b7a143dcad503045857bc49662f736437380ad45"}, + {file = "ruff-0.4.8-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:6ea874950daca5697309d976c9afba830d3bf0ed66887481d6bca1673fc5b66a"}, + {file = "ruff-0.4.8-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:fc95aac2943ddf360376be9aa3107c8cf9640083940a8c5bd824be692d2216dc"}, + {file = "ruff-0.4.8-py3-none-musllinux_1_2_i686.whl", hash = "sha256:384154a1c3f4bf537bac69f33720957ee49ac8d484bfc91720cc94172026ceed"}, + {file = "ruff-0.4.8-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:e9d5ce97cacc99878aa0d084c626a15cd21e6b3d53fd6f9112b7fc485918e1fa"}, + {file = "ruff-0.4.8-py3-none-win32.whl", hash = "sha256:6d795d7639212c2dfd01991259460101c22aabf420d9b943f153ab9d9706e6a9"}, + {file = "ruff-0.4.8-py3-none-win_amd64.whl", hash = "sha256:e14a3a095d07560a9d6769a72f781d73259655919d9b396c650fc98a8157555d"}, + {file = "ruff-0.4.8-py3-none-win_arm64.whl", hash = "sha256:14019a06dbe29b608f6b7cbcec300e3170a8d86efaddb7b23405cb7f7dcaf780"}, + {file = "ruff-0.4.8.tar.gz", hash = "sha256:16d717b1d57b2e2fd68bd0bf80fb43931b79d05a7131aa477d66fc40fbd86268"}, ] [[package]] @@ -3928,33 +3928,33 @@ files = [ [[package]] name = "tornado" -version = "6.4" +version = "6.4.1" description = "Tornado is a Python web framework and asynchronous networking library, originally developed at FriendFeed." optional = false -python-versions = ">= 3.8" +python-versions = ">=3.8" files = [ - {file = "tornado-6.4-cp38-abi3-macosx_10_9_universal2.whl", hash = "sha256:02ccefc7d8211e5a7f9e8bc3f9e5b0ad6262ba2fbb683a6443ecc804e5224ce0"}, - {file = "tornado-6.4-cp38-abi3-macosx_10_9_x86_64.whl", hash = "sha256:27787de946a9cffd63ce5814c33f734c627a87072ec7eed71f7fc4417bb16263"}, - {file = "tornado-6.4-cp38-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f7894c581ecdcf91666a0912f18ce5e757213999e183ebfc2c3fdbf4d5bd764e"}, - {file = "tornado-6.4-cp38-abi3-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:e43bc2e5370a6a8e413e1e1cd0c91bedc5bd62a74a532371042a18ef19e10579"}, - {file = "tornado-6.4-cp38-abi3-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f0251554cdd50b4b44362f73ad5ba7126fc5b2c2895cc62b14a1c2d7ea32f212"}, - {file = "tornado-6.4-cp38-abi3-musllinux_1_1_aarch64.whl", hash = "sha256:fd03192e287fbd0899dd8f81c6fb9cbbc69194d2074b38f384cb6fa72b80e9c2"}, - {file = "tornado-6.4-cp38-abi3-musllinux_1_1_i686.whl", hash = "sha256:88b84956273fbd73420e6d4b8d5ccbe913c65d31351b4c004ae362eba06e1f78"}, - {file = "tornado-6.4-cp38-abi3-musllinux_1_1_x86_64.whl", hash = "sha256:71ddfc23a0e03ef2df1c1397d859868d158c8276a0603b96cf86892bff58149f"}, - {file = "tornado-6.4-cp38-abi3-win32.whl", hash = "sha256:6f8a6c77900f5ae93d8b4ae1196472d0ccc2775cc1dfdc9e7727889145c45052"}, - {file = "tornado-6.4-cp38-abi3-win_amd64.whl", hash = "sha256:10aeaa8006333433da48dec9fe417877f8bcc21f48dda8d661ae79da357b2a63"}, - {file = "tornado-6.4.tar.gz", hash = "sha256:72291fa6e6bc84e626589f1c29d90a5a6d593ef5ae68052ee2ef000dfd273dee"}, + {file = "tornado-6.4.1-cp38-abi3-macosx_10_9_universal2.whl", hash = "sha256:163b0aafc8e23d8cdc3c9dfb24c5368af84a81e3364745ccb4427669bf84aec8"}, + {file = "tornado-6.4.1-cp38-abi3-macosx_10_9_x86_64.whl", hash = "sha256:6d5ce3437e18a2b66fbadb183c1d3364fb03f2be71299e7d10dbeeb69f4b2a14"}, + {file = "tornado-6.4.1-cp38-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e2e20b9113cd7293f164dc46fffb13535266e713cdb87bd2d15ddb336e96cfc4"}, + {file = "tornado-6.4.1-cp38-abi3-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:8ae50a504a740365267b2a8d1a90c9fbc86b780a39170feca9bcc1787ff80842"}, + {file = "tornado-6.4.1-cp38-abi3-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:613bf4ddf5c7a95509218b149b555621497a6cc0d46ac341b30bd9ec19eac7f3"}, + {file = "tornado-6.4.1-cp38-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:25486eb223babe3eed4b8aecbac33b37e3dd6d776bc730ca14e1bf93888b979f"}, + {file = "tornado-6.4.1-cp38-abi3-musllinux_1_2_i686.whl", hash = "sha256:454db8a7ecfcf2ff6042dde58404164d969b6f5d58b926da15e6b23817950fc4"}, + {file = "tornado-6.4.1-cp38-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:a02a08cc7a9314b006f653ce40483b9b3c12cda222d6a46d4ac63bb6c9057698"}, + {file = "tornado-6.4.1-cp38-abi3-win32.whl", hash = "sha256:d9a566c40b89757c9aa8e6f032bcdb8ca8795d7c1a9762910c722b1635c9de4d"}, + {file = "tornado-6.4.1-cp38-abi3-win_amd64.whl", hash = "sha256:b24b8982ed444378d7f21d563f4180a2de31ced9d8d84443907a0a64da2072e7"}, + {file = "tornado-6.4.1.tar.gz", hash = "sha256:92d3ab53183d8c50f8204a51e6f91d18a15d5ef261e84d452800d4ff6fc504e9"}, ] [[package]] name = "tox" -version = "4.15.0" +version = "4.15.1" description = "tox is a generic virtualenv management and test command line tool" optional = false python-versions = ">=3.8" files = [ - {file = "tox-4.15.0-py3-none-any.whl", hash = "sha256:300055f335d855b2ab1b12c5802de7f62a36d4fd53f30bd2835f6a201dda46ea"}, - {file = "tox-4.15.0.tar.gz", hash = "sha256:7a0beeef166fbe566f54f795b4906c31b428eddafc0102ac00d20998dd1933f6"}, + {file = "tox-4.15.1-py3-none-any.whl", hash = "sha256:f00a5dc4222b358e69694e47e3da0227ac41253509bca9f45aa8f012053e8d9d"}, + {file = "tox-4.15.1.tar.gz", hash = "sha256:53a092527d65e873e39213ebd4bd027a64623320b6b0326136384213f95b7076"}, ] [package.dependencies] @@ -4227,13 +4227,13 @@ test = ["pytest (>=6.0.0)", "setuptools (>=65)"] [[package]] name = "widgetsnbextension" -version = "4.0.10" +version = "4.0.11" description = "Jupyter interactive widgets for Jupyter Notebook" optional = false python-versions = ">=3.7" files = [ - {file = "widgetsnbextension-4.0.10-py3-none-any.whl", hash = "sha256:d37c3724ec32d8c48400a435ecfa7d3e259995201fbefa37163124a9fcb393cc"}, - {file = "widgetsnbextension-4.0.10.tar.gz", hash = "sha256:64196c5ff3b9a9183a8e699a4227fb0b7002f252c814098e66c4d1cd0644688f"}, + {file = "widgetsnbextension-4.0.11-py3-none-any.whl", hash = "sha256:55d4d6949d100e0d08b94948a42efc3ed6dfdc0e9468b2c4b128c9a2ce3a7a36"}, + {file = "widgetsnbextension-4.0.11.tar.gz", hash = "sha256:8b22a8f1910bfd188e596fe7fc05dcbd87e810c8a4ba010bdb3da86637398474"}, ] [[package]] @@ -4258,4 +4258,4 @@ desktop = ["opencv-python"] [metadata] lock-version = "2.0" python-versions = "^3.8" -content-hash = "ad8402ec1767f9427ab38bad7dab54b302a30f9e08b6489fad224c8481745b37" +content-hash = "e3d79f6c93041323b04c7b45e93bb3c4198b21889044004af8a0485a6145a207" diff --git a/pyproject.toml b/pyproject.toml index ff83f5fa..59d4176d 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,6 +1,6 @@ [tool.poetry] name = "supervision" -version = "0.21.0rc5" +version = "0.21.0" description = "A set of easy-to-use utils that will come in handy in any Computer Vision project" authors = ["Piotr Skalski "] maintainers = ["Piotr Skalski "] @@ -42,7 +42,7 @@ pyyaml = ">=5.3" defusedxml = "^0.7.1" opencv-python = { version = ">=4.5.5.64", optional = true } opencv-python-headless = ">=4.5.5.64" -requests = { version = ">=2.26.0,<=2.32.2", optional = true } +requests = { version = ">=2.26.0,<=2.32.3", optional = true } tqdm = { version = ">=4.62.3,<=4.66.4", optional = true } pillow = ">=9.4" diff --git a/supervision/__init__.py b/supervision/__init__.py index abe63390..4f28d49f 100644 --- a/supervision/__init__.py +++ b/supervision/__init__.py @@ -39,6 +39,13 @@ from supervision.dataset.utils import mask_to_rle, rle_to_mask from supervision.detection.annotate import BoxAnnotator from supervision.detection.core import Detections from supervision.detection.line_zone import LineZone, LineZoneAnnotator +from supervision.detection.lmm import LMM +from supervision.detection.overlap_filter import ( + OverlapFilter, + box_non_max_merge, + box_non_max_suppression, + mask_non_max_suppression, +) from supervision.detection.tools.csv_sink import CSVSink from supervision.detection.tools.inference_slicer import InferenceSlicer from supervision.detection.tools.json_sink import JSONSink @@ -46,14 +53,12 @@ from supervision.detection.tools.polygon_zone import PolygonZone, PolygonZoneAnn from supervision.detection.tools.smoother import DetectionsSmoother from supervision.detection.utils import ( box_iou_batch, - box_non_max_suppression, calculate_masks_centroids, clip_boxes, contains_holes, contains_multiple_segments, filter_polygons_by_area, mask_iou_batch, - mask_non_max_suppression, mask_to_polygons, mask_to_xyxy, move_boxes, diff --git a/supervision/detection/core.py b/supervision/detection/core.py index d939293f..37dde153 100644 --- a/supervision/detection/core.py +++ b/supervision/detection/core.py @@ -8,20 +8,24 @@ import numpy as np from supervision.config import CLASS_NAME_DATA_FIELD, ORIENTED_BOX_COORDINATES from supervision.detection.lmm import LMM, from_paligemma, validate_lmm_and_kwargs -from supervision.detection.utils import ( +from supervision.detection.overlap_filter import ( + box_non_max_merge, box_non_max_suppression, + mask_non_max_suppression, +) +from supervision.detection.utils import ( + box_iou_batch, calculate_masks_centroids, extract_ultralytics_masks, get_data_item, is_data_equal, - mask_non_max_suppression, mask_to_xyxy, merge_data, process_roboflow_result, xywh_to_xyxy, ) from supervision.geometry.core import Position -from supervision.utils.internal import deprecated +from supervision.utils.internal import deprecated, get_instance_variables from supervision.validators import validate_detections_fields @@ -874,6 +878,14 @@ class Detections: class_id=np.array([], dtype=int), ) + def is_empty(self) -> bool: + """ + Returns `True` if the `Detections` object is considered empty. + """ + empty_detections = Detections.empty() + empty_detections.data = self.data + return self == empty_detections + @classmethod def merge(cls, detections_list: List[Detections]) -> Detections: """ @@ -886,6 +898,10 @@ class Detections: For example, if merging Detections with 3 and 4 detected objects, this method will return a Detections with 7 objects (7 entries in `xyxy`, `mask`, etc). + !!! Note + + When merging, empty `Detections` objects are ignored. + Args: detections_list (List[Detections]): A list of Detections objects to merge. @@ -924,6 +940,10 @@ class Detections: array([0.1, 0.2, 0.3]) ``` """ + detections_list = [ + detections for detections in detections_list if not detections.is_empty() + ] + if len(detections_list) == 0: return Detections.empty() @@ -942,12 +962,13 @@ class Detections: def stack_or_none(name: str): if all(d.__getattribute__(name) is None for d in detections_list): return None - stack_list = [ - d.__getattribute__(name) - for d in detections_list - if d.__getattribute__(name) is not None - ] - return np.vstack(stack_list) if name == "mask" else np.hstack(stack_list) + if any(d.__getattribute__(name) is None for d in detections_list): + raise ValueError(f"All or none of the '{name}' fields must be None") + return ( + np.vstack([d.__getattribute__(name) for d in detections_list]) + if name == "mask" + else np.hstack([d.__getattribute__(name) for d in detections_list]) + ) mask = stack_or_none("mask") confidence = stack_or_none("confidence") @@ -1197,3 +1218,195 @@ class Detections: ) return self[indices] + + def with_nmm( + self, threshold: float = 0.5, class_agnostic: bool = False + ) -> Detections: + """ + Perform non-maximum merging on the current set of object detections. + + Args: + threshold (float, optional): The intersection-over-union threshold + to use for non-maximum merging. Defaults to 0.5. + class_agnostic (bool, optional): Whether to perform class-agnostic + non-maximum merging. If True, the class_id of each detection + will be ignored. Defaults to False. + + Returns: + Detections: A new Detections object containing the subset of detections + after non-maximum merging. + + Raises: + AssertionError: If `confidence` is None or `class_id` is None and + class_agnostic is False. + + ![non-max-merging](https://media.roboflow.com/supervision-docs/non-max-merging.png){ align=center width="800" } + """ # noqa: E501 // docs + if len(self) == 0: + return self + + assert ( + self.confidence is not None + ), "Detections confidence must be given for NMM to be executed." + + if class_agnostic: + predictions = np.hstack((self.xyxy, self.confidence.reshape(-1, 1))) + else: + assert self.class_id is not None, ( + "Detections class_id must be given for NMM to be executed. If you" + " intended to perform class agnostic NMM set class_agnostic=True." + ) + predictions = np.hstack( + ( + self.xyxy, + self.confidence.reshape(-1, 1), + self.class_id.reshape(-1, 1), + ) + ) + + merge_groups = box_non_max_merge( + predictions=predictions, iou_threshold=threshold + ) + + result = [] + for merge_group in merge_groups: + unmerged_detections = [self[i] for i in merge_group] + merged_detections = merge_inner_detections_objects( + unmerged_detections, threshold + ) + result.append(merged_detections) + + return Detections.merge(result) + + +def merge_inner_detection_object_pair( + detections_1: Detections, detections_2: Detections +) -> Detections: + """ + Merges two Detections object into a single Detections object. + Assumes each Detections contains exactly one object. + + A `winning` detection is determined based on the confidence score of the two + input detections. This winning detection is then used to specify which + `class_id`, `tracker_id`, and `data` to include in the merged Detections object. + + The resulting `confidence` of the merged object is calculated by the weighted + contribution of ea detection to the merged object. + The bounding boxes and masks of the two input detections are merged into a + single bounding box and mask, respectively. + + Args: + detections_1 (Detections): + The first Detections object + detections_2 (Detections): + The second Detections object + + Returns: + Detections: A new Detections object, with merged attributes. + + Raises: + ValueError: If the input Detections objects do not have exactly 1 detected + object. + + Example: + ```python + import cv2 + import supervision as sv + from inference import get_model + + image = cv2.imread() + model = get_model(model_id="yolov8s-640") + + result = model.infer(image)[0] + detections = sv.Detections.from_inference(result) + + merged_detections = merge_object_detection_pair( + detections[0], detections[1]) + ``` + """ + if len(detections_1) != 1 or len(detections_2) != 1: + raise ValueError("Both Detections should have exactly 1 detected object.") + + validate_fields_both_defined_or_none(detections_1, detections_2) + + xyxy_1 = detections_1.xyxy[0] + xyxy_2 = detections_2.xyxy[0] + if detections_1.confidence is None and detections_2.confidence is None: + merged_confidence = None + else: + detection_1_area = (xyxy_1[2] - xyxy_1[0]) * (xyxy_1[3] - xyxy_1[1]) + detections_2_area = (xyxy_2[2] - xyxy_2[0]) * (xyxy_2[3] - xyxy_2[1]) + merged_confidence = ( + detection_1_area * detections_1.confidence[0] + + detections_2_area * detections_2.confidence[0] + ) / (detection_1_area + detections_2_area) + merged_confidence = np.array([merged_confidence]) + + merged_x1, merged_y1 = np.minimum(xyxy_1[:2], xyxy_2[:2]) + merged_x2, merged_y2 = np.maximum(xyxy_1[2:], xyxy_2[2:]) + merged_xyxy = np.array([[merged_x1, merged_y1, merged_x2, merged_y2]]) + + if detections_1.mask is None and detections_2.mask is None: + merged_mask = None + else: + merged_mask = np.logical_or(detections_1.mask, detections_2.mask) + + if detections_1.confidence is None and detections_2.confidence is None: + winning_detection = detections_1 + elif detections_1.confidence[0] >= detections_2.confidence[0]: + winning_detection = detections_1 + else: + winning_detection = detections_2 + + return Detections( + xyxy=merged_xyxy, + mask=merged_mask, + confidence=merged_confidence, + class_id=winning_detection.class_id, + tracker_id=winning_detection.tracker_id, + data=winning_detection.data, + ) + + +def merge_inner_detections_objects( + detections: List[Detections], threshold=0.5 +) -> Detections: + """ + Given N detections each of length 1 (exactly one object inside), combine them into a + single detection object of length 1. The contained inner object will be the merged + result of all the input detections. + + For example, this lets you merge N boxes into one big box, N masks into one mask, + etc. + """ + detections_1 = detections[0] + for detections_2 in detections[1:]: + box_iou = box_iou_batch(detections_1.xyxy, detections_2.xyxy)[0] + if box_iou < threshold: + break + detections_1 = merge_inner_detection_object_pair(detections_1, detections_2) + return detections_1 + + +def validate_fields_both_defined_or_none( + detections_1: Detections, detections_2: Detections +) -> None: + """ + Verify that for each optional field in the Detections, both instances either have + the field set to None or both have it set to non-None values. + + `data` field is ignored. + + Raises: + ValueError: If one field is None and the other is not, for any of the fields. + """ + attributes = get_instance_variables(detections_1) + for attribute in attributes: + value_1 = getattr(detections_1, attribute) + value_2 = getattr(detections_2, attribute) + + if (value_1 is None) != (value_2 is None): + raise ValueError( + f"Field '{attribute}' should be consistently None or not None in both " + "Detections." + ) diff --git a/supervision/detection/line_zone.py b/supervision/detection/line_zone.py index cd4400e5..4255a0ad 100644 --- a/supervision/detection/line_zone.py +++ b/supervision/detection/line_zone.py @@ -5,6 +5,7 @@ import cv2 import numpy as np from supervision.detection.core import Detections +from supervision.detection.utils import cross_product from supervision.draw.color import Color from supervision.draw.utils import draw_text from supervision.geometry.core import Point, Position, Vector @@ -83,6 +84,8 @@ class LineZone: self.in_count: int = 0 self.out_count: int = 0 self.triggering_anchors = triggering_anchors + if not list(self.triggering_anchors): + raise ValueError("Triggering anchors cannot be empty.") @staticmethod def calculate_region_of_interest_limits(vector: Vector) -> Tuple[Vector, Vector]: @@ -158,28 +161,23 @@ class LineZone: ] ) + cross_products_1 = cross_product(all_anchors, self.limits[0]) + cross_products_2 = cross_product(all_anchors, self.limits[1]) + in_limits = (cross_products_1 > 0) == (cross_products_2 > 0) + in_limits = np.all(in_limits, axis=0) + + triggers = cross_product(all_anchors, self.vector) < 0 + has_any_left_trigger = np.any(triggers, axis=0) + has_any_right_trigger = np.any(~triggers, axis=0) + is_uniformly_triggered = ~(has_any_left_trigger & has_any_right_trigger) for i, tracker_id in enumerate(detections.tracker_id): - box_anchors = [Point(x=x, y=y) for x, y in all_anchors[:, i, :]] - - in_limits = all( - [ - self.is_point_in_limits(point=anchor, limits=self.limits) - for anchor in box_anchors - ] - ) - - if not in_limits: + if not in_limits[i]: continue - triggers = [ - self.vector.cross_product(point=anchor) < 0 for anchor in box_anchors - ] - - if len(set(triggers)) == 2: + if not is_uniformly_triggered[i]: continue - tracker_state = triggers[0] - + tracker_state = has_any_left_trigger[i] if tracker_id not in self.tracker_state: self.tracker_state[tracker_id] = tracker_state continue diff --git a/supervision/detection/lmm.py b/supervision/detection/lmm.py index 0278fc00..5f61db0a 100644 --- a/supervision/detection/lmm.py +++ b/supervision/detection/lmm.py @@ -41,7 +41,7 @@ def from_paligemma( ) -> Tuple[np.ndarray, Optional[np.ndarray], np.ndarray]: w, h = resolution_wh pattern = re.compile( - r"(?) ([\w\s]+)" + r"(?) ([\w\s\-]+)" ) matches = pattern.findall(result) matches = np.array(matches) if matches else np.empty((0, 5)) diff --git a/supervision/detection/overlap_filter.py b/supervision/detection/overlap_filter.py new file mode 100644 index 00000000..ab4408d1 --- /dev/null +++ b/supervision/detection/overlap_filter.py @@ -0,0 +1,263 @@ +from enum import Enum +from typing import List, Union + +import numpy as np +import numpy.typing as npt + +from supervision.detection.utils import box_iou_batch, mask_iou_batch + + +def resize_masks(masks: np.ndarray, max_dimension: int = 640) -> np.ndarray: + """ + Resize all masks in the array to have a maximum dimension of max_dimension, + maintaining aspect ratio. + + Args: + masks (np.ndarray): 3D array of binary masks with shape (N, H, W). + max_dimension (int): The maximum dimension for the resized masks. + + Returns: + np.ndarray: Array of resized masks. + """ + max_height = np.max(masks.shape[1]) + max_width = np.max(masks.shape[2]) + scale = min(max_dimension / max_height, max_dimension / max_width) + + new_height = int(scale * max_height) + new_width = int(scale * max_width) + + x = np.linspace(0, max_width - 1, new_width).astype(int) + y = np.linspace(0, max_height - 1, new_height).astype(int) + xv, yv = np.meshgrid(x, y) + + resized_masks = masks[:, yv, xv] + + resized_masks = resized_masks.reshape(masks.shape[0], new_height, new_width) + return resized_masks + + +def mask_non_max_suppression( + predictions: np.ndarray, + masks: np.ndarray, + iou_threshold: float = 0.5, + mask_dimension: int = 640, +) -> np.ndarray: + """ + Perform Non-Maximum Suppression (NMS) on segmentation predictions. + + Args: + predictions (np.ndarray): A 2D array of object detection predictions in + the format of `(x_min, y_min, x_max, y_max, score)` + or `(x_min, y_min, x_max, y_max, score, class)`. Shape: `(N, 5)` or + `(N, 6)`, where N is the number of predictions. + masks (np.ndarray): A 3D array of binary masks corresponding to the predictions. + Shape: `(N, H, W)`, where N is the number of predictions, and H, W are the + dimensions of each mask. + iou_threshold (float, optional): The intersection-over-union threshold + to use for non-maximum suppression. + mask_dimension (int, optional): The dimension to which the masks should be + resized before computing IOU values. Defaults to 640. + + Returns: + np.ndarray: A boolean array indicating which predictions to keep after + non-maximum suppression. + + Raises: + AssertionError: If `iou_threshold` is not within the closed + range from `0` to `1`. + """ + assert 0 <= iou_threshold <= 1, ( + "Value of `iou_threshold` must be in the closed range from 0 to 1, " + f"{iou_threshold} given." + ) + rows, columns = predictions.shape + + if columns == 5: + predictions = np.c_[predictions, np.zeros(rows)] + + sort_index = predictions[:, 4].argsort()[::-1] + predictions = predictions[sort_index] + masks = masks[sort_index] + masks_resized = resize_masks(masks, mask_dimension) + ious = mask_iou_batch(masks_resized, masks_resized) + categories = predictions[:, 5] + + keep = np.ones(rows, dtype=bool) + for i in range(rows): + if keep[i]: + condition = (ious[i] > iou_threshold) & (categories[i] == categories) + keep[i + 1 :] = np.where(condition[i + 1 :], False, keep[i + 1 :]) + + return keep[sort_index.argsort()] + + +def box_non_max_suppression( + predictions: np.ndarray, iou_threshold: float = 0.5 +) -> np.ndarray: + """ + Perform Non-Maximum Suppression (NMS) on object detection predictions. + + Args: + predictions (np.ndarray): An array of object detection predictions in + the format of `(x_min, y_min, x_max, y_max, score)` + or `(x_min, y_min, x_max, y_max, score, class)`. + iou_threshold (float, optional): The intersection-over-union threshold + to use for non-maximum suppression. + + Returns: + np.ndarray: A boolean array indicating which predictions to keep after n + on-maximum suppression. + + Raises: + AssertionError: If `iou_threshold` is not within the + closed range from `0` to `1`. + """ + assert 0 <= iou_threshold <= 1, ( + "Value of `iou_threshold` must be in the closed range from 0 to 1, " + f"{iou_threshold} given." + ) + rows, columns = predictions.shape + + # add column #5 - category filled with zeros for agnostic nms + if columns == 5: + predictions = np.c_[predictions, np.zeros(rows)] + + # sort predictions column #4 - score + sort_index = np.flip(predictions[:, 4].argsort()) + predictions = predictions[sort_index] + + boxes = predictions[:, :4] + categories = predictions[:, 5] + ious = box_iou_batch(boxes, boxes) + ious = ious - np.eye(rows) + + keep = np.ones(rows, dtype=bool) + + for index, (iou, category) in enumerate(zip(ious, categories)): + if not keep[index]: + continue + + # drop detections with iou > iou_threshold and + # same category as current detections + condition = (iou > iou_threshold) & (categories == category) + keep = keep & ~condition + + return keep[sort_index.argsort()] + + +def group_overlapping_boxes( + predictions: npt.NDArray[np.float64], iou_threshold: float = 0.5 +) -> List[List[int]]: + """ + Apply greedy version of non-maximum merging to avoid detecting too many + overlapping bounding boxes for a given object. + + Args: + predictions (npt.NDArray[np.float64]): An array of shape `(n, 5)` containing + the bounding boxes coordinates in format `[x1, y1, x2, y2]` + and the confidence scores. + iou_threshold (float, optional): The intersection-over-union threshold + to use for non-maximum suppression. Defaults to 0.5. + + Returns: + List[List[int]]: Groups of prediction indices be merged. + Each group may have 1 or more elements. + """ + merge_groups: List[List[int]] = [] + + scores = predictions[:, 4] + order = scores.argsort() + + while len(order) > 0: + idx = int(order[-1]) + + order = order[:-1] + if len(order) == 0: + merge_groups.append([idx]) + break + + merge_candidate = np.expand_dims(predictions[idx], axis=0) + ious = box_iou_batch(predictions[order][:, :4], merge_candidate[:, :4]) + ious = ious.flatten() + + above_threshold = ious >= iou_threshold + merge_group = [idx] + np.flip(order[above_threshold]).tolist() + merge_groups.append(merge_group) + order = order[~above_threshold] + return merge_groups + + +def box_non_max_merge( + predictions: npt.NDArray[np.float64], + iou_threshold: float = 0.5, +) -> List[List[int]]: + """ + Apply greedy version of non-maximum merging per category to avoid detecting + too many overlapping bounding boxes for a given object. + + Args: + predictions (npt.NDArray[np.float64]): An array of shape `(n, 5)` or `(n, 6)` + containing the bounding boxes coordinates in format `[x1, y1, x2, y2]`, + the confidence scores and class_ids. Omit class_id column to allow + detections of different classes to be merged. + iou_threshold (float, optional): The intersection-over-union threshold + to use for non-maximum suppression. Defaults to 0.5. + + Returns: + List[List[int]]: Groups of prediction indices be merged. + Each group may have 1 or more elements. + """ + if predictions.shape[1] == 5: + return group_overlapping_boxes(predictions, iou_threshold) + + category_ids = predictions[:, 5] + merge_groups = [] + for category_id in np.unique(category_ids): + curr_indices = np.where(category_ids == category_id)[0] + merge_class_groups = group_overlapping_boxes( + predictions[curr_indices], iou_threshold + ) + + for merge_class_group in merge_class_groups: + merge_groups.append(curr_indices[merge_class_group].tolist()) + + for merge_group in merge_groups: + if len(merge_group) == 0: + raise ValueError( + f"Empty group detected when non-max-merging " + f"detections: {merge_groups}" + ) + return merge_groups + + +class OverlapFilter(Enum): + """ + Enum specifying the strategy for filtering overlapping detections. + + Attributes: + NONE: Do not filter detections based on overlap. + NON_MAX_SUPPRESSION: Filter detections using non-max suppression. This means, + detections that overlap by more than a set threshold will be discarded, + except for the one with the highest confidence. + NON_MAX_MERGE: Merge detections with non-max merging. This means, + detections that overlap by more than a set threshold will be merged + into a single detection. + """ + + NONE = "none" + NON_MAX_SUPPRESSION = "non_max_suppression" + NON_MAX_MERGE = "non_max_merge" + + +def validate_overlap_filter( + strategy: Union[OverlapFilter, str], +) -> OverlapFilter: + if isinstance(strategy, str): + try: + strategy = OverlapFilter(strategy.lower()) + except ValueError: + raise ValueError( + f"Invalid strategy value: {strategy}. Must be one of " + f"{[e.value for e in OverlapFilter]}" + ) + return strategy diff --git a/supervision/detection/tools/inference_slicer.py b/supervision/detection/tools/inference_slicer.py index 82551434..134361bd 100644 --- a/supervision/detection/tools/inference_slicer.py +++ b/supervision/detection/tools/inference_slicer.py @@ -1,11 +1,14 @@ +import warnings from concurrent.futures import ThreadPoolExecutor, as_completed -from typing import Callable, Optional, Tuple +from typing import Callable, Optional, Tuple, Union import numpy as np from supervision.detection.core import Detections +from supervision.detection.overlap_filter import OverlapFilter, validate_overlap_filter from supervision.detection.utils import move_boxes, move_masks from supervision.utils.image import crop_image +from supervision.utils.internal import SupervisionWarnings def move_detections( @@ -50,8 +53,10 @@ class InferenceSlicer: `(width, height)`. overlap_ratio_wh (Tuple[float, float]): Overlap ratio between consecutive slices in the format `(width_ratio, height_ratio)`. - iou_threshold (Optional[float]): Intersection over Union (IoU) threshold - used for non-max suppression. + overlap_filter_strategy (Union[OverlapFilter, str]): Strategy for + filtering or merging overlapping detections in slices. + iou_threshold (float): Intersection over Union (IoU) threshold + used when filtering by overlap. callback (Callable): A function that performs inference on a given image slice and returns detections. thread_workers (int): Number of threads for parallel execution. @@ -68,12 +73,18 @@ class InferenceSlicer: callback: Callable[[np.ndarray], Detections], slice_wh: Tuple[int, int] = (320, 320), overlap_ratio_wh: Tuple[float, float] = (0.2, 0.2), - iou_threshold: Optional[float] = 0.5, + overlap_filter_strategy: Union[ + OverlapFilter, str + ] = OverlapFilter.NON_MAX_SUPPRESSION, + iou_threshold: float = 0.5, thread_workers: int = 1, ): + overlap_filter_strategy = validate_overlap_filter(overlap_filter_strategy) + self.slice_wh = slice_wh self.overlap_ratio_wh = overlap_ratio_wh self.iou_threshold = iou_threshold + self.overlap_filter_strategy = overlap_filter_strategy self.callback = callback self.thread_workers = thread_workers @@ -104,7 +115,10 @@ class InferenceSlicer: result = model(image_slice)[0] return sv.Detections.from_ultralytics(result) - slicer = sv.InferenceSlicer(callback = callback) + slicer = sv.InferenceSlicer( + callback=callback, + overlap_filter_strategy=sv.OverlapFilter.NON_MAX_SUPPRESSION, + ) detections = slicer(image) ``` @@ -124,9 +138,19 @@ class InferenceSlicer: for future in as_completed(futures): detections_list.append(future.result()) - return Detections.merge(detections_list=detections_list).with_nms( - threshold=self.iou_threshold - ) + merged = Detections.merge(detections_list=detections_list) + if self.overlap_filter_strategy == OverlapFilter.NONE: + return merged + elif self.overlap_filter_strategy == OverlapFilter.NON_MAX_SUPPRESSION: + return merged.with_nms(threshold=self.iou_threshold) + elif self.overlap_filter_strategy == OverlapFilter.NON_MAX_MERGE: + return merged.with_nmm(threshold=self.iou_threshold) + else: + warnings.warn( + f"Invalid overlap filter strategy: {self.overlap_filter_strategy}", + category=SupervisionWarnings, + ) + return merged def _run_callback(self, image, offset) -> Detections: """ diff --git a/supervision/detection/tools/polygon_zone.py b/supervision/detection/tools/polygon_zone.py index a1997212..f1c48f94 100644 --- a/supervision/detection/tools/polygon_zone.py +++ b/supervision/detection/tools/polygon_zone.py @@ -54,6 +54,8 @@ class PolygonZone: self.polygon = polygon.astype(int) self.triggering_anchors = triggering_anchors + if not list(self.triggering_anchors): + raise ValueError("Triggering anchors cannot be empty.") self.current_count = 0 diff --git a/supervision/detection/utils.py b/supervision/detection/utils.py index aac0d627..5b92aedd 100644 --- a/supervision/detection/utils.py +++ b/supervision/detection/utils.py @@ -6,6 +6,7 @@ import numpy as np import numpy.typing as npt from supervision.config import CLASS_NAME_DATA_FIELD +from supervision.geometry.core import Vector MIN_POLYGON_POINT_COUNT = 3 @@ -139,144 +140,6 @@ def mask_iou_batch( return np.vstack(ious) -def resize_masks(masks: np.ndarray, max_dimension: int = 640) -> np.ndarray: - """ - Resize all masks in the array to have a maximum dimension of max_dimension, - maintaining aspect ratio. - - Args: - masks (np.ndarray): 3D array of binary masks with shape (N, H, W). - max_dimension (int): The maximum dimension for the resized masks. - - Returns: - np.ndarray: Array of resized masks. - """ - max_height = np.max(masks.shape[1]) - max_width = np.max(masks.shape[2]) - scale = min(max_dimension / max_height, max_dimension / max_width) - - new_height = int(scale * max_height) - new_width = int(scale * max_width) - - x = np.linspace(0, max_width - 1, new_width).astype(int) - y = np.linspace(0, max_height - 1, new_height).astype(int) - xv, yv = np.meshgrid(x, y) - - resized_masks = masks[:, yv, xv] - - resized_masks = resized_masks.reshape(masks.shape[0], new_height, new_width) - return resized_masks - - -def mask_non_max_suppression( - predictions: np.ndarray, - masks: np.ndarray, - iou_threshold: float = 0.5, - mask_dimension: int = 640, -) -> np.ndarray: - """ - Perform Non-Maximum Suppression (NMS) on segmentation predictions. - - Args: - predictions (np.ndarray): A 2D array of object detection predictions in - the format of `(x_min, y_min, x_max, y_max, score)` - or `(x_min, y_min, x_max, y_max, score, class)`. Shape: `(N, 5)` or - `(N, 6)`, where N is the number of predictions. - masks (np.ndarray): A 3D array of binary masks corresponding to the predictions. - Shape: `(N, H, W)`, where N is the number of predictions, and H, W are the - dimensions of each mask. - iou_threshold (float, optional): The intersection-over-union threshold - to use for non-maximum suppression. - mask_dimension (int, optional): The dimension to which the masks should be - resized before computing IOU values. Defaults to 640. - - Returns: - np.ndarray: A boolean array indicating which predictions to keep after - non-maximum suppression. - - Raises: - AssertionError: If `iou_threshold` is not within the closed - range from `0` to `1`. - """ - assert 0 <= iou_threshold <= 1, ( - "Value of `iou_threshold` must be in the closed range from 0 to 1, " - f"{iou_threshold} given." - ) - rows, columns = predictions.shape - - if columns == 5: - predictions = np.c_[predictions, np.zeros(rows)] - - sort_index = predictions[:, 4].argsort()[::-1] - predictions = predictions[sort_index] - masks = masks[sort_index] - masks_resized = resize_masks(masks, mask_dimension) - ious = mask_iou_batch(masks_resized, masks_resized) - categories = predictions[:, 5] - - keep = np.ones(rows, dtype=bool) - for i in range(rows): - if keep[i]: - condition = (ious[i] > iou_threshold) & (categories[i] == categories) - keep[i + 1 :] = np.where(condition[i + 1 :], False, keep[i + 1 :]) - - return keep[sort_index.argsort()] - - -def box_non_max_suppression( - predictions: np.ndarray, iou_threshold: float = 0.5 -) -> np.ndarray: - """ - Perform Non-Maximum Suppression (NMS) on object detection predictions. - - Args: - predictions (np.ndarray): An array of object detection predictions in - the format of `(x_min, y_min, x_max, y_max, score)` - or `(x_min, y_min, x_max, y_max, score, class)`. - iou_threshold (float, optional): The intersection-over-union threshold - to use for non-maximum suppression. - - Returns: - np.ndarray: A boolean array indicating which predictions to keep after n - on-maximum suppression. - - Raises: - AssertionError: If `iou_threshold` is not within the - closed range from `0` to `1`. - """ - assert 0 <= iou_threshold <= 1, ( - "Value of `iou_threshold` must be in the closed range from 0 to 1, " - f"{iou_threshold} given." - ) - rows, columns = predictions.shape - - # add column #5 - category filled with zeros for agnostic nms - if columns == 5: - predictions = np.c_[predictions, np.zeros(rows)] - - # sort predictions column #4 - score - sort_index = np.flip(predictions[:, 4].argsort()) - predictions = predictions[sort_index] - - boxes = predictions[:, :4] - categories = predictions[:, 5] - ious = box_iou_batch(boxes, boxes) - ious = ious - np.eye(rows) - - keep = np.ones(rows, dtype=bool) - - for index, (iou, category) in enumerate(zip(ious, categories)): - if not keep[index]: - continue - - # drop detections with iou > iou_threshold and - # same category as current detections - condition = (iou > iou_threshold) & (categories == category) - keep = keep & ~condition - - return keep[sort_index.argsort()] - - def clip_boxes(xyxy: np.ndarray, resolution_wh: Tuple[int, int]) -> np.ndarray: """ Clips bounding boxes coordinates to fit within the frame resolution. @@ -292,6 +155,25 @@ def clip_boxes(xyxy: np.ndarray, resolution_wh: Tuple[int, int]) -> np.ndarray: np.ndarray: A numpy array of shape `(N, 4)` where each row corresponds to a bounding box with coordinates clipped to fit within the frame resolution. + + Examples: + ```python + import numpy as np + import supervision as sv + + xyxy = np.array([ + [10, 20, 300, 200], + [15, 25, 350, 450], + [-10, -20, 30, 40] + ]) + + sv.clip_boxes(xyxy=xyxy, resolution_wh=(320, 240)) + # array([ + # [ 10, 20, 300, 200], + # [ 15, 25, 320, 240], + # [ 0, 0, 30, 40] + # ]) + ``` """ result = np.copy(xyxy) width, height = resolution_wh @@ -318,6 +200,23 @@ def pad_boxes(xyxy: np.ndarray, px: int, py: Optional[int] = None) -> np.ndarray np.ndarray: A numpy array of shape `(N, 4)` where each row corresponds to a bounding box with coordinates padded according to the provided padding values. + + Examples: + ```python + import numpy as np + import supervision as sv + + xyxy = np.array([ + [10, 20, 30, 40], + [15, 25, 35, 45] + ]) + + sv.pad_boxes(xyxy=xyxy, px=5, py=10) + # array([ + # [ 5, 10, 35, 50], + # [10, 15, 40, 55] + # ]) + ``` """ if py is None: py = px @@ -349,7 +248,7 @@ def mask_to_xyxy(masks: np.ndarray) -> np.ndarray: `(x_min, y_min, x_max, y_max)` for each mask """ n = masks.shape[0] - bboxes = np.zeros((n, 4), dtype=int) + xyxy = np.zeros((n, 4), dtype=int) for i, mask in enumerate(masks): rows, cols = np.where(mask) @@ -357,9 +256,9 @@ def mask_to_xyxy(masks: np.ndarray) -> np.ndarray: if len(rows) > 0 and len(cols) > 0: x_min, x_max = np.min(cols), np.max(cols) y_min, y_max = np.min(rows), np.max(rows) - bboxes[i, :] = [x_min, y_min, x_max, y_max] + xyxy[i, :] = [x_min, y_min, x_max, y_max] - return bboxes + return xyxy def mask_to_polygons(mask: np.ndarray) -> List[np.ndarray]: @@ -595,16 +494,18 @@ def process_roboflow_result( return xyxy, confidence, class_id, masks, tracker_id, data -def move_boxes(xyxy: np.ndarray, offset: np.ndarray) -> np.ndarray: +def move_boxes( + xyxy: npt.NDArray[np.float64], offset: npt.NDArray[np.int32] +) -> npt.NDArray[np.float64]: """ Parameters: - xyxy (np.ndarray): An array of shape `(n, 4)` containing the bounding boxes - coordinates in format `[x1, y1, x2, y2]` + xyxy (npt.NDArray[np.float64]): An array of shape `(n, 4)` containing the + bounding boxes coordinates in format `[x1, y1, x2, y2]` offset (np.array): An array of shape `(2,)` containing offset values in format is `[dx, dy]`. Returns: - np.ndarray: Repositioned bounding boxes. + npt.NDArray[np.float64]: Repositioned bounding boxes. Examples: ```python @@ -628,24 +529,25 @@ def move_boxes(xyxy: np.ndarray, offset: np.ndarray) -> np.ndarray: def move_masks( - masks: np.ndarray, - offset: np.ndarray, - resolution_wh: Tuple[int, int] = None, -) -> np.ndarray: + masks: npt.NDArray[np.bool_], + offset: npt.NDArray[np.int32], + resolution_wh: Tuple[int, int], +) -> npt.NDArray[np.bool_]: """ Offset the masks in an array by the specified (x, y) amount. Args: - masks (np.ndarray): A 3D array of binary masks corresponding to the predictions. - Shape: `(N, H, W)`, where N is the number of predictions, and H, W are the - dimensions of each mask. - offset (np.ndarray): An array of shape `(2,)` containing non-negative int values - `[dx, dy]`. + masks (npt.NDArray[np.bool_]): A 3D array of binary masks corresponding to the + predictions. Shape: `(N, H, W)`, where N is the number of predictions, and + H, W are the dimensions of each mask. + offset (npt.NDArray[np.int32]): An array of shape `(2,)` containing non-negative + int values `[dx, dy]`. resolution_wh (Tuple[int, int]): The width and height of the desired mask resolution. Returns: - (np.ndarray) repositioned masks, optionally padded to the specified shape. + (npt.NDArray[np.bool_]) repositioned masks, optionally padded to the specified + shape. """ if offset[0] < 0 or offset[1] < 0: @@ -661,19 +563,21 @@ def move_masks( return mask_array -def scale_boxes(xyxy: np.ndarray, factor: float) -> np.ndarray: +def scale_boxes( + xyxy: npt.NDArray[np.float64], factor: float +) -> npt.NDArray[np.float64]: """ Scale the dimensions of bounding boxes. Parameters: - xyxy (np.ndarray): An array of shape `(n, 4)` containing the bounding boxes - coordinates in format `[x1, y1, x2, y2]` + xyxy (npt.NDArray[np.float64]): An array of shape `(n, 4)` containing the + bounding boxes coordinates in format `[x1, y1, x2, y2]` factor (float): A float value representing the factor by which the box dimensions are scaled. A factor greater than 1 enlarges the boxes, while a factor less than 1 shrinks them. Returns: - np.ndarray: Scaled bounding boxes. + npt.NDArray[np.float64]: Scaled bounding boxes. Examples: ```python @@ -685,7 +589,7 @@ def scale_boxes(xyxy: np.ndarray, factor: float) -> np.ndarray: [30, 30, 40, 40] ]) - scaled_bb = sv.scale_boxes(xyxy=xyxy, factor=1.5) + sv.scale_boxes(xyxy=xyxy, factor=1.5) # array([ # [ 7.5, 7.5, 22.5, 22.5], # [27.5, 27.5, 42.5, 42.5] @@ -743,19 +647,19 @@ def is_data_equal(data_a: Dict[str, np.ndarray], data_b: Dict[str, np.ndarray]) def merge_data( - data_list: List[Dict[str, Union[np.ndarray, List]]], -) -> Dict[str, Union[np.ndarray, List]]: + data_list: List[Dict[str, Union[npt.NDArray[np.generic], List]]], +) -> Dict[str, Union[npt.NDArray[np.generic], List]]: """ Merges the data payloads of a list of Detections instances. Args: data_list: The data payloads of the Detections instances. Each data payload is a dictionary with the same keys, and the values are either lists or - np.ndarray. + npt.NDArray[np.generic]. Returns: A single data payload containing the merged data, preserving the original data - types (list or np.ndarray). + types (list or npt.NDArray[np.generic]). Raises: ValueError: If data values within a single object have different lengths or if @@ -764,6 +668,10 @@ def merge_data( if not data_list: return {} + all_keys_sets = [set(data.keys()) for data in data_list] + if not all(keys_set == all_keys_sets[0] for keys_set in all_keys_sets): + raise ValueError("All data dictionaries must have the same keys to merge.") + for data in data_list: lengths = [len(value) for value in data.values()] if len(set(lengths)) > 1: @@ -771,21 +679,7 @@ def merge_data( "All data values within a single object must have equal length." ) - keys_by_data = [set(data.keys()) for data in data_list] - keys_by_data = [keys for keys in keys_by_data if len(keys) > 0] - if not keys_by_data: - return {} - - common_keys = set.intersection(*keys_by_data) - all_keys = set.union(*keys_by_data) - if common_keys != all_keys: - raise ValueError( - f"All sv.Detections.data dictionaries must have the same keys. Common " - f"keys: {common_keys}, but some dictionaries have additional keys: " - f"{all_keys.difference(common_keys)}." - ) - - merged_data = {key: [] for key in all_keys} + merged_data = {key: [] for key in all_keys_sets[0]} for data in data_list: for key in data: merged_data[key].append(data[key]) @@ -966,3 +860,20 @@ def contains_multiple_segments( mask_uint8, labels, connectivity=connectivity ) return number_of_labels > 2 + + +def cross_product(anchors: np.ndarray, vector: Vector) -> np.ndarray: + """ + Get array of cross products of each anchor with a vector. + Args: + anchors: Array of anchors of shape (number of anchors, detections, 2) + vector: Vector to calculate cross product with + + Returns: + Array of cross products of shape (number of anchors, detections) + """ + vector_at_zero = np.array( + [vector.end.x - vector.start.x, vector.end.y - vector.start.y] + ) + vector_start = np.array([vector.start.x, vector.start.y]) + return np.cross(vector_at_zero, anchors - vector_start) diff --git a/supervision/utils/internal.py b/supervision/utils/internal.py index 978a1448..072c03b7 100644 --- a/supervision/utils/internal.py +++ b/supervision/utils/internal.py @@ -1,7 +1,8 @@ import functools +import inspect import os import warnings -from typing import Callable +from typing import Any, Callable, Set class SupervisionWarnings(Warning): @@ -141,3 +142,42 @@ class classproperty(property): The result of calling the function stored in 'fget' with 'owner_cls'. """ return self.fget(owner_cls) + + +def get_instance_variables(instance: Any, include_properties=False) -> Set[str]: + """ + Get the public variables of a class instance. + + Args: + instance (Any): The instance of a class + include_properties (bool): Whether to include properties in the result + + Usage: + ```python + detections = Detections(xyxy=np.array([1,2,3,4])) + variables = get_class_variables(detections) + # ["xyxy", "mask", "confidence", ..., "data"] + ``` + """ + if isinstance(instance, type): + raise ValueError("Only class instances are supported, not classes.") + + fields = set( + ( + name + for name, val in inspect.getmembers(instance) + if not callable(val) and not name.startswith("_") + ) + ) + + if not include_properties: + properties = set( + ( + name + for name, val in inspect.getmembers(instance.__class__) + if isinstance(val, property) + ) + ) + fields -= properties + + return fields diff --git a/test/detection/test_core.py b/test/detection/test_core.py index 12f3de28..300d6dfe 100644 --- a/test/detection/test_core.py +++ b/test/detection/test_core.py @@ -5,7 +5,7 @@ from typing import List, Optional, Union import numpy as np import pytest -from supervision.detection.core import Detections +from supervision.detection.core import Detections, merge_inner_detection_object_pair from supervision.geometry.core import Position PREDICTIONS = np.array( @@ -245,7 +245,6 @@ def test_getitem( TEST_DET_1_2, DoesNotRaise(), ), # Fields with same keys - # Fields and empty ( [TEST_DET_1, Detections.empty()], TEST_DET_1, @@ -264,9 +263,9 @@ def test_getitem( TEST_DET_1, TEST_DET_NONE, ], - TEST_DET_1, - DoesNotRaise(), - ), # Single detection and None fields (+ missing Dict keys) + None, + pytest.raises(ValueError), + ), # Empty detection, but not Detections.empty() # Errors: Non-zero-length differently defined keys & data ( [TEST_DET_1, TEST_DET_DIFFERENT_FIELDS], @@ -278,6 +277,22 @@ def test_getitem( None, pytest.raises(ValueError), ), # Non-empty detections with different data keys + ( + [ + mock_detections( + xyxy=[[10, 10, 20, 20]], + class_id=[1], + mask=[np.zeros((4, 4), dtype=bool)], + ), + Detections.empty(), + ], + mock_detections( + xyxy=[[10, 10, 20, 20]], + class_id=[1], + mask=[np.zeros((4, 4), dtype=bool)], + ), + DoesNotRaise(), + ), # Segmentation + Empty ], ) def test_merge( @@ -421,3 +436,172 @@ def test_equal( detections_a: Detections, detections_b: Detections, expected_result: bool ) -> None: assert (detections_a == detections_b) == expected_result + + +@pytest.mark.parametrize( + "detection_1, detection_2, expected_result, exception", + [ + ( + mock_detections( + xyxy=[[10, 10, 30, 30]], + ), + mock_detections( + xyxy=[[10, 10, 30, 30]], + ), + mock_detections( + xyxy=[[10, 10, 30, 30]], + ), + DoesNotRaise(), + ), # Merge with self + ( + mock_detections( + xyxy=[[10, 10, 30, 30]], + ), + Detections.empty(), + None, + pytest.raises(ValueError), + ), # merge with empty: error + ( + mock_detections( + xyxy=[[10, 10, 30, 30]], + ), + mock_detections( + xyxy=[[10, 10, 30, 30], [40, 40, 60, 60]], + ), + None, + pytest.raises(ValueError), + ), # merge with 2+ objects: error + ( + mock_detections( + xyxy=[[10, 10, 30, 30]], + confidence=[0.1], + class_id=[1], + mask=[np.array([[1, 1, 0], [1, 1, 0], [0, 0, 0]], dtype=bool)], + tracker_id=[1], + data={"key_1": [1]}, + ), + mock_detections( + xyxy=[[20, 20, 40, 40]], + confidence=[0.1], + class_id=[2], + mask=[np.array([[0, 0, 0], [0, 1, 1], [0, 1, 1]], dtype=bool)], + tracker_id=[2], + data={"key_2": [2]}, + ), + mock_detections( + xyxy=[[10, 10, 40, 40]], + confidence=[0.1], + class_id=[1], + mask=[np.array([[1, 1, 0], [1, 1, 1], [0, 1, 1]], dtype=bool)], + tracker_id=[1], + data={"key_1": [1]}, + ), + DoesNotRaise(), + ), # Same confidence - merge box & mask, tie-break to detection_1 + ( + mock_detections( + xyxy=[[0, 0, 20, 20]], + confidence=[0.1], + class_id=[1], + mask=[np.array([[1, 1, 0], [1, 1, 0], [0, 0, 0]], dtype=bool)], + tracker_id=[1], + data={"key_1": [1]}, + ), + mock_detections( + xyxy=[[10, 10, 50, 50]], + confidence=[0.2], + class_id=[2], + mask=[np.array([[0, 0, 0], [0, 1, 1], [0, 1, 1]], dtype=bool)], + tracker_id=[2], + data={"key_2": [2]}, + ), + mock_detections( + xyxy=[[0, 0, 50, 50]], + confidence=[(1 * 0.1 + 4 * 0.2) / 5], + class_id=[2], + mask=[np.array([[1, 1, 0], [1, 1, 1], [0, 1, 1]], dtype=bool)], + tracker_id=[2], + data={"key_2": [2]}, + ), + DoesNotRaise(), + ), # Different confidence, different area + ( + mock_detections( + xyxy=[[10, 10, 30, 30]], + confidence=None, + class_id=[1], + mask=[np.array([[1, 1, 0], [1, 1, 0], [0, 0, 0]], dtype=bool)], + tracker_id=[1], + data={"key_1": [1]}, + ), + mock_detections( + xyxy=[[20, 20, 40, 40]], + confidence=None, + class_id=[2], + mask=[np.array([[0, 0, 0], [0, 1, 1], [0, 1, 1]], dtype=bool)], + tracker_id=[2], + data={"key_2": [2]}, + ), + mock_detections( + xyxy=[[10, 10, 40, 40]], + confidence=None, + class_id=[1], + mask=[np.array([[1, 1, 0], [1, 1, 1], [0, 1, 1]], dtype=bool)], + tracker_id=[1], + data={"key_1": [1]}, + ), + DoesNotRaise(), + ), # No confidence at all + ( + mock_detections( + xyxy=[[0, 0, 20, 20]], + confidence=None, + ), + mock_detections( + xyxy=[[10, 10, 30, 30]], + confidence=[0.2], + ), + None, + pytest.raises(ValueError), + ), # confidence: None + [x] + ( + mock_detections( + xyxy=[[0, 0, 20, 20]], + mask=[np.array([[1, 1, 0], [1, 1, 0], [0, 0, 0]], dtype=bool)], + ), + mock_detections( + xyxy=[[10, 10, 30, 30]], + mask=None, + ), + None, + pytest.raises(ValueError), + ), # mask: None + [x] + ( + mock_detections(xyxy=[[0, 0, 20, 20]], tracker_id=[1]), + mock_detections( + xyxy=[[10, 10, 30, 30]], + tracker_id=None, + ), + None, + pytest.raises(ValueError), + ), # tracker_id: None + [] + ( + mock_detections(xyxy=[[0, 0, 20, 20]], class_id=[1]), + mock_detections( + xyxy=[[10, 10, 30, 30]], + class_id=None, + ), + None, + pytest.raises(ValueError), + ), # class_id: None + [] + ], +) +def test_merge_inner_detection_object_pair( + detection_1: Detections, + detection_2: Detections, + expected_result: Optional[Detections], + exception: Exception, +): + with exception: + result = merge_inner_detection_object_pair(detection_1, detection_2) + assert result == expected_result diff --git a/test/detection/test_line_counter.py b/test/detection/test_line_counter.py index 73780414..66118e97 100644 --- a/test/detection/test_line_counter.py +++ b/test/detection/test_line_counter.py @@ -1,10 +1,11 @@ from contextlib import ExitStack as DoesNotRaise -from typing import Optional, Tuple +from test.test_utils import mock_detections +from typing import List, Optional, Tuple import pytest from supervision import LineZone -from supervision.geometry.core import Point, Vector +from supervision.geometry.core import Point, Position, Vector @pytest.mark.parametrize( @@ -70,3 +71,409 @@ def test_calculate_region_of_interest_limits( with exception: result = LineZone.calculate_region_of_interest_limits(vector=vector) assert result == expected_result + + +@pytest.mark.parametrize( + "vector, xyxy_sequence, expected_crossed_in, expected_crossed_out", + [ + ( # Vertical line, simple crossing + Vector(Point(0, 0), Point(0, 10)), + [ + [4, 4, 6, 6], + [4 - 10, 4, 6 - 10, 6], + [4, 4, 6, 6], + [4 - 10, 4, 6 - 10, 6], + ], + [False, False, True, False], + [False, True, False, True], + ), + ( # Vertical line reversed, simple crossing + Vector(Point(0, 10), Point(0, 0)), + [ + [4, 4, 6, 6], + [4 - 10, 4, 6 - 10, 6], + [4, 4, 6, 6], + [4 - 10, 4, 6 - 10, 6], + ], + [False, True, False, True], + [False, False, True, False], + ), + ( # Horizontal line, simple crossing + Vector(Point(0, 0), Point(10, 0)), + [ + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + ], + [False, True, False, True], + [False, False, True, False], + ), + ( # Horizontal line reversed, simple crossing + Vector(Point(10, 0), Point(0, 0)), + [ + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + ], + [False, False, True, False], + [False, True, False, True], + ), + ( # Diagonal line, simple crossing + Vector(Point(5, 0), Point(0, 5)), + [ + [0, 0, 2, 2], + [0 + 10, 0 + 10, 2 + 10, 2 + 10], + [0, 0, 2, 2], + [0 + 10, 0 + 10, 2 + 10, 2 + 10], + ], + [False, True, False, True], + [False, False, True, False], + ), + ( # Crossing beside - right side + Vector(Point(0, 0), Point(10, 0)), + [ + [20, 4, 24, 6], + [20, 4 - 10, 24, 6 - 10], + [20, 4, 24, 6], + [20, 4 - 10, 24, 6 - 10], + ], + [False, False, False, False], + [False, False, False, False], + ), + ( # Horizontal line, simple crossing, far away + Vector(Point(0, 0), Point(10, 0)), + [ + [4, 1e32, 6, 1e32 + 2], + [4, -1e32, 6, -1e32 + 2], + [4, 1e32, 6, 1e32 + 2], + [4, -1e32, 6, -1e32 + 2], + ], + [False, True, False, True], + [False, False, True, False], + ), + ( # Crossing beside - left side + Vector(Point(0, 0), Point(10, 0)), + [ + [-20, 4, -24, 6], + [-20, 4 - 10, -24, 6 - 10], + [-20, 4, -24, 6], + [-20, 4 - 10, -24, 6 - 10], + ], + [False, False, False, False], + [False, False, False, False], + ), + ( # Move above + Vector(Point(0, 0), Point(10, 0)), + [ + [-4, 4, -2, 6], + [-4 + 20, 4, -2 + 20, 6], + [-4, 4, -2, 6], + [-4 + 20, 4, -2 + 20, 6], + ], + [False, False, False, False], + [False, False, False, False], + ), + ( # Move below + Vector(Point(0, 0), Point(10, 0)), + [ + [-4, -6, -2, -4], + [-4 + 20, -6, -2 + 20, -4], + [-4, -6, -2, -4], + [-4 + 20, -6, -2 + 20, -4], + ], + [False, False, False, False], + [False, False, False, False], + ), + ( # Move into line partway + Vector(Point(0, 0), Point(10, 0)), + [ + [4, 4, 6, 6], + [4 + 5, 4, 6 + 5, 6], + [4, 4, 6, 6], + [4 + 5, 4, 6 + 5, 6], + ], + [False, False, False, False], + [False, False, False, False], + ), + ( # V-shaped crossing from outside limits - not supported. + Vector(Point(0, 0), Point(10, 0)), + [[-3, 6, -1, 8], [4, -6, 6, -4], [11, 6, 13, 8]], + [False, False, False], + [False, False, False], + ), + ( # Diagonal movement, from within limits to outside - not supported + Vector(Point(0, 0), Point(10, 0)), + [[4, 1, 6, 3], [11, 1 - 20, 13, 3 - 20]], + [False, False], + [False, False], + ), + ( # Diagonal movement, from outside limits to within - not supported + Vector(Point(0, 0), Point(10, 0)), + [ + [11, 21, 13, 23], + [4, -3, 6, -1], + ], + [False, False], + [False, False], + ), + ( # Diagonal crossing, from outside to outside limits - not supported. + Vector(Point(0, 0), Point(10, 0)), + [ + [-4, 4, -2, 8], + [-4 + 16, -4, -2 + 16, -6], + [-4, 4, -2, 8], + [-4 + 16, -4, -2 + 16, -6], + ], + [False, False, False, False], + [False, False, False, False], + ), + ], +) +def test_line_zone_one_detection_default_anchors( + vector: Vector, + xyxy_sequence: List[List[float]], + expected_crossed_in: List[bool], + expected_crossed_out: List[bool], +) -> None: + line_zone = LineZone(start=vector.start, end=vector.end) + + crossed_in_list = [] + crossed_out_list = [] + for i, bbox in enumerate(xyxy_sequence): + detections = mock_detections( + xyxy=[bbox], + tracker_id=[0], + ) + crossed_in, crossed_out = line_zone.trigger(detections) + crossed_in_list.append(crossed_in[0]) + crossed_out_list.append(crossed_out[0]) + + assert ( + crossed_in_list == expected_crossed_in + ), f"expected {expected_crossed_in}, got {crossed_in_list}" + assert ( + crossed_out_list == expected_crossed_out + ), f"expected {expected_crossed_out}, got {crossed_out_list}" + + +@pytest.mark.parametrize( + "vector, xyxy_sequence, triggering_anchors, expected_crossed_in, " + "expected_crossed_out", + [ + ( # Scrape line, left side, corner anchors + Vector(Point(0, 0), Point(10, 0)), + [ + [-2, 4, 2, 6], + [-2, 4 - 10, 2, 6 - 10], + [-2, 4, 2, 6], + [-2, 4 - 10, 2, 6 - 10], + ], + [ + Position.TOP_LEFT, + Position.BOTTOM_LEFT, + Position.TOP_RIGHT, + Position.BOTTOM_RIGHT, + ], + [False, False, False, False], + [False, False, False, False], + ), + ( # Scrape line, left side, right anchors + Vector(Point(0, 0), Point(10, 0)), + [ + [-2, 4, 2, 6], + [-2, 4 - 10, 2, 6 - 10], + [-2, 4, 2, 6], + [-2, 4 - 10, 2, 6 - 10], + ], + [Position.TOP_RIGHT, Position.BOTTOM_RIGHT], + [False, True, False, True], + [False, False, True, False], + ), + ( # Scrape line, left side, center anchor (along line point) + Vector(Point(0, 0), Point(10, 0)), + [ + [-2, 4, 2, 6], + [-2, 4 - 10, 2, 6 - 10], + [-2, 4, 2, 6], + [-2, 4 - 10, 2, 6 - 10], + ], + [Position.CENTER], + [False, True, False, True], + [False, False, True, False], + ), + ( # Scrape line, right side, corner anchors + Vector(Point(0, 0), Point(10, 0)), + [ + [8, 4, 12, 6], + [8, 4 - 10, 12, 6 - 10], + [8, 4, 12, 6], + [8, 4 - 10, 12, 6 - 10], + ], + [ + Position.TOP_LEFT, + Position.BOTTOM_LEFT, + Position.TOP_RIGHT, + Position.BOTTOM_RIGHT, + ], + [False, False, False, False], + [False, False, False, False], + ), + ( # Scrape line, right side, left anchors + Vector(Point(0, 0), Point(10, 0)), + [ + [8, 4, 12, 6], + [8, 4 - 10, 12, 6 - 10], + [8, 4, 12, 6], + [8, 4 - 10, 12, 6 - 10], + ], + [Position.TOP_LEFT, Position.BOTTOM_LEFT], + [False, True, False, True], + [False, False, True, False], + ), + ( # Scrape line, right side, center anchor (along line point) + Vector(Point(0, 0), Point(10, 0)), + [ + [8, 4, 12, 6], + [8, 4 - 10, 12, 6 - 10], + [8, 4, 12, 6], + [8, 4 - 10, 12, 6 - 10], + ], + [Position.CENTER], + [False, True, False, True], + [False, False, True, False], + ), + ( # Simple crossing, one anchor + Vector(Point(0, 0), Point(10, 0)), + [ + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + ], + [Position.CENTER], + [False, True, False, True], + [False, False, True, False], + ), + ( # Simple crossing, all box anchors + Vector(Point(0, 0), Point(10, 0)), + [ + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + [4, 4, 6, 6], + [4, 4 - 10, 6, 6 - 10], + ], + [ + Position.CENTER, + Position.CENTER_LEFT, + Position.CENTER_RIGHT, + Position.TOP_CENTER, + Position.TOP_LEFT, + Position.TOP_RIGHT, + Position.BOTTOM_LEFT, + Position.BOTTOM_CENTER, + Position.BOTTOM_RIGHT, + ], + [False, True, False, True], + [False, False, True, False], + ), + ], +) +def test_line_zone_one_detection( + vector: Vector, + xyxy_sequence: List[List[float]], + triggering_anchors: List[Position], + expected_crossed_in: List[bool], + expected_crossed_out: List[bool], +) -> None: + line_zone = LineZone( + start=vector.start, end=vector.end, triggering_anchors=triggering_anchors + ) + + crossed_in_list = [] + crossed_out_list = [] + for i, bbox in enumerate(xyxy_sequence): + detections = mock_detections( + xyxy=[bbox], + tracker_id=[0], + ) + crossed_in, crossed_out = line_zone.trigger(detections) + crossed_in_list.append(crossed_in[0]) + crossed_out_list.append(crossed_out[0]) + + assert ( + crossed_in_list == expected_crossed_in + ), f"expected {expected_crossed_in}, got {crossed_in_list}" + assert ( + crossed_out_list == expected_crossed_out + ), f"expected {expected_crossed_out}, got {crossed_out_list}" + + +@pytest.mark.parametrize( + "vector, xyxy_sequence, anchors, expected_crossed_in, " + "expected_crossed_out, exception", + [ + ( # One stays, one crosses + Vector(Point(0, 0), Point(10, 0)), + [ + [[4, 4, 6, 6], [4, 4, 6, 6]], + [[4, 4, 6, 6], [4, 4 - 10, 6, 6 - 10]], + [[4, 4, 6, 6], [4, 4, 6, 6]], + [[4, 4, 6, 6], [4, 4 - 10, 6, 6 - 10]], + ], + [ + Position.TOP_LEFT, + Position.TOP_RIGHT, + Position.BOTTOM_LEFT, + Position.BOTTOM_RIGHT, + ], + [[False, False], [False, True], [False, False], [False, True]], + [[False, False], [False, False], [False, True], [False, False]], + DoesNotRaise(), + ), + ( # Both cross at the same time + Vector(Point(0, 0), Point(10, 0)), + [ + [[4, 4, 6, 6], [4, 4, 6, 6]], + [[4, 4 - 10, 6, 6 - 10], [4, 4 - 10, 6, 6 - 10]], + [[4, 4, 6, 6], [4, 4, 6, 6]], + [[4, 4 - 10, 6, 6 - 10], [4, 4 - 10, 6, 6 - 10]], + ], + [ + Position.TOP_LEFT, + Position.TOP_RIGHT, + Position.BOTTOM_LEFT, + Position.BOTTOM_RIGHT, + ], + [[False, False], [True, True], [False, False], [True, True]], + [[False, False], [False, False], [True, True], [False, False]], + DoesNotRaise(), + ), + ], +) +def test_line_zone_multiple_detections( + vector: Vector, + xyxy_sequence: List[List[List[float]]], + anchors: List[Position], + expected_crossed_in: List[List[bool]], + expected_crossed_out: List[List[bool]], + exception: Exception, +) -> None: + with exception: + line_zone = LineZone( + start=vector.start, end=vector.end, triggering_anchors=anchors + ) + crossed_in_list = [] + crossed_out_list = [] + for bboxes in xyxy_sequence: + detections = mock_detections( + xyxy=bboxes, + tracker_id=[i for i in range(0, len(bboxes))], + ) + crossed_in, crossed_out = line_zone.trigger(detections) + crossed_in_list.append(list(crossed_in)) + crossed_out_list.append(list(crossed_out)) + + assert crossed_in_list == expected_crossed_in + assert crossed_out_list == expected_crossed_out diff --git a/test/detection/test_lmm.py b/test/detection/test_lmm.py index 129aa44b..4448d8db 100644 --- a/test/detection/test_lmm.py +++ b/test/detection/test_lmm.py @@ -76,7 +76,27 @@ from supervision.detection.lmm import from_paligemma None, np.array(["black cat"]).astype(np.dtype("U")), ), - ), # correct response; no classes + ), # correct response; class name with space; no classes + ( + " black-cat", + (1000, 1000), + None, + ( + np.array([[250.0, 250.0, 750.0, 750.0]]), + None, + np.array(["black-cat"]).astype(np.dtype("U")), + ), + ), # correct response; class name with hyphen; no classes + ( + " black_cat", + (1000, 1000), + None, + ( + np.array([[250.0, 250.0, 750.0, 750.0]]), + None, + np.array(["black_cat"]).astype(np.dtype("U")), + ), + ), # correct response; class name with underscore; no classes ( " cat ;", (1000, 1000), diff --git a/test/detection/test_overlap_filter.py b/test/detection/test_overlap_filter.py new file mode 100644 index 00000000..f628c30f --- /dev/null +++ b/test/detection/test_overlap_filter.py @@ -0,0 +1,449 @@ +from contextlib import ExitStack as DoesNotRaise +from typing import List, Optional + +import numpy as np +import pytest + +from supervision.detection.overlap_filter import ( + box_non_max_suppression, + group_overlapping_boxes, + mask_non_max_suppression, +) + + +@pytest.mark.parametrize( + "predictions, iou_threshold, expected_result, exception", + [ + ( + np.empty(shape=(0, 5), dtype=float), + 0.5, + [], + DoesNotRaise(), + ), + ( + np.array([[0, 0, 10, 10, 1.0]]), + 0.5, + [[0]], + DoesNotRaise(), + ), + ( + np.array([[0, 0, 10, 10, 1.0], [0, 0, 9, 9, 1.0]]), + 0.5, + [[1, 0]], + DoesNotRaise(), + ), # High overlap, tie-break to second det + ( + np.array([[0, 0, 10, 10, 1.0], [0, 0, 9, 9, 0.99]]), + 0.5, + [[0, 1]], + DoesNotRaise(), + ), # High overlap, merge to high confidence + ( + np.array([[0, 0, 10, 10, 0.99], [0, 0, 9, 9, 1.0]]), + 0.5, + [[1, 0]], + DoesNotRaise(), + ), # (test symmetry) High overlap, merge to high confidence + ( + np.array([[0, 0, 10, 10, 0.90], [0, 0, 9, 9, 1.0]]), + 0.5, + [[1, 0]], + DoesNotRaise(), + ), # (test symmetry) High overlap, merge to high confidence + ( + np.array([[0, 0, 10, 10, 1.0], [0, 0, 9, 9, 1.0]]), + 1.0, + [[1], [0]], + DoesNotRaise(), + ), # High IOU required + ( + np.array([[0, 0, 10, 10, 1.0], [0, 0, 9, 9, 1.0]]), + 0.0, + [[1, 0]], + DoesNotRaise(), + ), # No IOU required + ( + np.array([[0, 0, 10, 10, 1.0], [0, 0, 5, 5, 0.9]]), + 0.25, + [[0, 1]], + DoesNotRaise(), + ), # Below IOU requirement + ( + np.array([[0, 0, 10, 10, 1.0], [0, 0, 5, 5, 0.9]]), + 0.26, + [[0], [1]], + DoesNotRaise(), + ), # Above IOU requirement + ( + np.array([[0, 0, 10, 10, 1.0], [0, 0, 9, 9, 1.0], [0, 0, 8, 8, 1.0]]), + 0.5, + [[2, 1, 0]], + DoesNotRaise(), + ), # 3 boxes + ( + np.array( + [ + [0, 0, 10, 10, 1.0], + [0, 0, 9, 9, 1.0], + [5, 5, 10, 10, 1.0], + [6, 6, 10, 10, 1.0], + [9, 9, 10, 10, 1.0], + ] + ), + 0.5, + [[4], [3, 2], [1, 0]], + DoesNotRaise(), + ), # 5 boxes, 2 merges, 1 separate + ( + np.array( + [ + [0, 0, 2, 1, 1.0], + [1, 0, 3, 1, 1.0], + [2, 0, 4, 1, 1.0], + [3, 0, 5, 1, 1.0], + [4, 0, 6, 1, 1.0], + ] + ), + 0.33, + [[4, 3], [2, 1], [0]], + DoesNotRaise(), + ), # sequential merge, half overlap + ( + np.array( + [ + [0, 0, 2, 1, 0.9], + [1, 0, 3, 1, 0.9], + [2, 0, 4, 1, 1.0], + [3, 0, 5, 1, 0.9], + [4, 0, 6, 1, 0.9], + ] + ), + 0.33, + [[2, 3, 1], [4], [0]], + DoesNotRaise(), + ), # confidence + ], +) +def test_group_overlapping_boxes( + predictions: np.ndarray, + iou_threshold: float, + expected_result: List[List[int]], + exception: Exception, +) -> None: + with exception: + result = group_overlapping_boxes( + predictions=predictions, iou_threshold=iou_threshold + ) + + assert result == expected_result + + +@pytest.mark.parametrize( + "predictions, iou_threshold, expected_result, exception", + [ + ( + np.empty(shape=(0, 5)), + 0.5, + np.array([]), + DoesNotRaise(), + ), # single box with no category + ( + np.array([[10.0, 10.0, 40.0, 40.0, 0.8]]), + 0.5, + np.array([True]), + DoesNotRaise(), + ), # single box with no category + ( + np.array([[10.0, 10.0, 40.0, 40.0, 0.8, 0]]), + 0.5, + np.array([True]), + DoesNotRaise(), + ), # single box with category + ( + np.array( + [ + [10.0, 10.0, 40.0, 40.0, 0.8], + [15.0, 15.0, 40.0, 40.0, 0.9], + ] + ), + 0.5, + np.array([False, True]), + DoesNotRaise(), + ), # two boxes with no category + ( + np.array( + [ + [10.0, 10.0, 40.0, 40.0, 0.8, 0], + [15.0, 15.0, 40.0, 40.0, 0.9, 1], + ] + ), + 0.5, + np.array([True, True]), + DoesNotRaise(), + ), # two boxes with different category + ( + np.array( + [ + [10.0, 10.0, 40.0, 40.0, 0.8, 0], + [15.0, 15.0, 40.0, 40.0, 0.9, 0], + ] + ), + 0.5, + np.array([False, True]), + DoesNotRaise(), + ), # two boxes with same category + ( + np.array( + [ + [0.0, 0.0, 30.0, 40.0, 0.8], + [5.0, 5.0, 35.0, 45.0, 0.9], + [10.0, 10.0, 40.0, 50.0, 0.85], + ] + ), + 0.5, + np.array([False, True, False]), + DoesNotRaise(), + ), # three boxes with no category + ( + np.array( + [ + [0.0, 0.0, 30.0, 40.0, 0.8, 0], + [5.0, 5.0, 35.0, 45.0, 0.9, 1], + [10.0, 10.0, 40.0, 50.0, 0.85, 2], + ] + ), + 0.5, + np.array([True, True, True]), + DoesNotRaise(), + ), # three boxes with same category + ( + np.array( + [ + [0.0, 0.0, 30.0, 40.0, 0.8, 0], + [5.0, 5.0, 35.0, 45.0, 0.9, 0], + [10.0, 10.0, 40.0, 50.0, 0.85, 1], + ] + ), + 0.5, + np.array([False, True, True]), + DoesNotRaise(), + ), # three boxes with different category + ], +) +def test_box_non_max_suppression( + predictions: np.ndarray, + iou_threshold: float, + expected_result: Optional[np.ndarray], + exception: Exception, +) -> None: + with exception: + result = box_non_max_suppression( + predictions=predictions, iou_threshold=iou_threshold + ) + assert np.array_equal(result, expected_result) + + +@pytest.mark.parametrize( + "predictions, masks, iou_threshold, expected_result, exception", + [ + ( + np.empty((0, 6)), + np.empty((0, 5, 5)), + 0.5, + np.array([]), + DoesNotRaise(), + ), # empty predictions and masks + ( + np.array([[0, 0, 0, 0, 0.8]]), + np.array( + [ + [ + [False, False, False, False, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, False, False, False, False], + ] + ] + ), + 0.5, + np.array([True]), + DoesNotRaise(), + ), # single mask with no category + ( + np.array([[0, 0, 0, 0, 0.8, 0]]), + np.array( + [ + [ + [False, False, False, False, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, False, False, False, False], + ] + ] + ), + 0.5, + np.array([True]), + DoesNotRaise(), + ), # single mask with category + ( + np.array([[0, 0, 0, 0, 0.8], [0, 0, 0, 0, 0.9]]), + np.array( + [ + [ + [False, False, False, False, False], + [False, True, True, False, False], + [False, True, True, False, False], + [False, False, False, False, False], + [False, False, False, False, False], + ], + [ + [False, False, False, False, False], + [False, False, False, False, False], + [False, False, False, True, True], + [False, False, False, True, True], + [False, False, False, False, False], + ], + ] + ), + 0.5, + np.array([True, True]), + DoesNotRaise(), + ), # two masks non-overlapping with no category + ( + np.array([[0, 0, 0, 0, 0.8], [0, 0, 0, 0, 0.9]]), + np.array( + [ + [ + [False, False, False, False, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, False, False, False, False], + ], + [ + [False, False, False, False, False], + [False, False, True, True, True], + [False, False, True, True, True], + [False, False, True, True, True], + [False, False, False, False, False], + ], + ] + ), + 0.4, + np.array([False, True]), + DoesNotRaise(), + ), # two masks partially overlapping with no category + ( + np.array([[0, 0, 0, 0, 0.8, 0], [0, 0, 0, 0, 0.9, 1]]), + np.array( + [ + [ + [False, False, False, False, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, True, True, True, False], + [False, False, False, False, False], + ], + [ + [False, False, False, False, False], + [False, False, True, True, True], + [False, False, True, True, True], + [False, False, True, True, True], + [False, False, False, False, False], + ], + ] + ), + 0.5, + np.array([True, True]), + DoesNotRaise(), + ), # two masks partially overlapping with different category + ( + np.array( + [ + [0, 0, 0, 0, 0.8], + [0, 0, 0, 0, 0.85], + [0, 0, 0, 0, 0.9], + ] + ), + np.array( + [ + [ + [False, False, False, False, False], + [False, True, True, False, False], + [False, True, True, False, False], + [False, False, False, False, False], + [False, False, False, False, False], + ], + [ + [False, False, False, False, False], + [False, True, True, False, False], + [False, True, True, False, False], + [False, False, False, False, False], + [False, False, False, False, False], + ], + [ + [False, False, False, False, False], + [False, False, False, True, True], + [False, False, False, True, True], + [False, False, False, False, False], + [False, False, False, False, False], + ], + ] + ), + 0.5, + np.array([False, True, True]), + DoesNotRaise(), + ), # three masks with no category + ( + np.array( + [ + [0, 0, 0, 0, 0.8, 0], + [0, 0, 0, 0, 0.85, 1], + [0, 0, 0, 0, 0.9, 2], + ] + ), + np.array( + [ + [ + [False, False, False, False, False], + [False, True, True, False, False], + [False, True, True, False, False], + [False, False, False, False, False], + [False, False, False, False, False], + ], + [ + [False, False, False, False, False], + [False, True, True, False, False], + [False, True, True, False, False], + [False, True, True, False, False], + [False, False, False, False, False], + ], + [ + [False, False, False, False, False], + [False, True, True, False, False], + [False, True, True, False, False], + [False, False, False, False, False], + [False, False, False, False, False], + ], + ] + ), + 0.5, + np.array([True, True, True]), + DoesNotRaise(), + ), # three masks with different category + ], +) +def test_mask_non_max_suppression( + predictions: np.ndarray, + masks: np.ndarray, + iou_threshold: float, + expected_result: Optional[np.ndarray], + exception: Exception, +) -> None: + with exception: + result = mask_non_max_suppression( + predictions=predictions, masks=masks, iou_threshold=iou_threshold + ) + assert np.array_equal(result, expected_result) diff --git a/test/detection/test_polygonzone.py b/test/detection/test_polygonzone.py index 1a86a45b..ed899615 100644 --- a/test/detection/test_polygonzone.py +++ b/test/detection/test_polygonzone.py @@ -92,3 +92,19 @@ def test_polygon_zone_trigger( with exception: in_zone = polygon_zone.trigger(detections) assert np.all(in_zone == expected_results) + + +@pytest.mark.parametrize( + "polygon, triggering_anchors, exception", + [ + (POLYGON, [sv.Position.CENTER], DoesNotRaise()), + ( + POLYGON, + [], + pytest.raises(ValueError), + ), + ], +) +def test_polygon_zone_initialization(polygon, triggering_anchors, exception): + with exception: + sv.PolygonZone(polygon, FRAME_RESOLUTION, triggering_anchors=triggering_anchors) diff --git a/test/detection/test_utils.py b/test/detection/test_utils.py index 6a2070da..20c818e6 100644 --- a/test/detection/test_utils.py +++ b/test/detection/test_utils.py @@ -7,14 +7,12 @@ import pytest from supervision.config import CLASS_NAME_DATA_FIELD from supervision.detection.utils import ( - box_non_max_suppression, calculate_masks_centroids, clip_boxes, contains_holes, contains_multiple_segments, filter_polygons_by_area, get_data_item, - mask_non_max_suppression, merge_data, move_boxes, process_roboflow_result, @@ -25,317 +23,6 @@ TEST_MASK = np.zeros((1, 1000, 1000), dtype=bool) TEST_MASK[:, 300:351, 200:251] = True -@pytest.mark.parametrize( - "predictions, iou_threshold, expected_result, exception", - [ - ( - np.empty(shape=(0, 5)), - 0.5, - np.array([]), - DoesNotRaise(), - ), # single box with no category - ( - np.array([[10.0, 10.0, 40.0, 40.0, 0.8]]), - 0.5, - np.array([True]), - DoesNotRaise(), - ), # single box with no category - ( - np.array([[10.0, 10.0, 40.0, 40.0, 0.8, 0]]), - 0.5, - np.array([True]), - DoesNotRaise(), - ), # single box with category - ( - np.array( - [ - [10.0, 10.0, 40.0, 40.0, 0.8], - [15.0, 15.0, 40.0, 40.0, 0.9], - ] - ), - 0.5, - np.array([False, True]), - DoesNotRaise(), - ), # two boxes with no category - ( - np.array( - [ - [10.0, 10.0, 40.0, 40.0, 0.8, 0], - [15.0, 15.0, 40.0, 40.0, 0.9, 1], - ] - ), - 0.5, - np.array([True, True]), - DoesNotRaise(), - ), # two boxes with different category - ( - np.array( - [ - [10.0, 10.0, 40.0, 40.0, 0.8, 0], - [15.0, 15.0, 40.0, 40.0, 0.9, 0], - ] - ), - 0.5, - np.array([False, True]), - DoesNotRaise(), - ), # two boxes with same category - ( - np.array( - [ - [0.0, 0.0, 30.0, 40.0, 0.8], - [5.0, 5.0, 35.0, 45.0, 0.9], - [10.0, 10.0, 40.0, 50.0, 0.85], - ] - ), - 0.5, - np.array([False, True, False]), - DoesNotRaise(), - ), # three boxes with no category - ( - np.array( - [ - [0.0, 0.0, 30.0, 40.0, 0.8, 0], - [5.0, 5.0, 35.0, 45.0, 0.9, 1], - [10.0, 10.0, 40.0, 50.0, 0.85, 2], - ] - ), - 0.5, - np.array([True, True, True]), - DoesNotRaise(), - ), # three boxes with same category - ( - np.array( - [ - [0.0, 0.0, 30.0, 40.0, 0.8, 0], - [5.0, 5.0, 35.0, 45.0, 0.9, 0], - [10.0, 10.0, 40.0, 50.0, 0.85, 1], - ] - ), - 0.5, - np.array([False, True, True]), - DoesNotRaise(), - ), # three boxes with different category - ], -) -def test_box_non_max_suppression( - predictions: np.ndarray, - iou_threshold: float, - expected_result: Optional[np.ndarray], - exception: Exception, -) -> None: - with exception: - result = box_non_max_suppression( - predictions=predictions, iou_threshold=iou_threshold - ) - assert np.array_equal(result, expected_result) - - -@pytest.mark.parametrize( - "predictions, masks, iou_threshold, expected_result, exception", - [ - ( - np.empty((0, 6)), - np.empty((0, 5, 5)), - 0.5, - np.array([]), - DoesNotRaise(), - ), # empty predictions and masks - ( - np.array([[0, 0, 0, 0, 0.8]]), - np.array( - [ - [ - [False, False, False, False, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, False, False, False, False], - ] - ] - ), - 0.5, - np.array([True]), - DoesNotRaise(), - ), # single mask with no category - ( - np.array([[0, 0, 0, 0, 0.8, 0]]), - np.array( - [ - [ - [False, False, False, False, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, False, False, False, False], - ] - ] - ), - 0.5, - np.array([True]), - DoesNotRaise(), - ), # single mask with category - ( - np.array([[0, 0, 0, 0, 0.8], [0, 0, 0, 0, 0.9]]), - np.array( - [ - [ - [False, False, False, False, False], - [False, True, True, False, False], - [False, True, True, False, False], - [False, False, False, False, False], - [False, False, False, False, False], - ], - [ - [False, False, False, False, False], - [False, False, False, False, False], - [False, False, False, True, True], - [False, False, False, True, True], - [False, False, False, False, False], - ], - ] - ), - 0.5, - np.array([True, True]), - DoesNotRaise(), - ), # two masks non-overlapping with no category - ( - np.array([[0, 0, 0, 0, 0.8], [0, 0, 0, 0, 0.9]]), - np.array( - [ - [ - [False, False, False, False, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, False, False, False, False], - ], - [ - [False, False, False, False, False], - [False, False, True, True, True], - [False, False, True, True, True], - [False, False, True, True, True], - [False, False, False, False, False], - ], - ] - ), - 0.4, - np.array([False, True]), - DoesNotRaise(), - ), # two masks partially overlapping with no category - ( - np.array([[0, 0, 0, 0, 0.8, 0], [0, 0, 0, 0, 0.9, 1]]), - np.array( - [ - [ - [False, False, False, False, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, True, True, True, False], - [False, False, False, False, False], - ], - [ - [False, False, False, False, False], - [False, False, True, True, True], - [False, False, True, True, True], - [False, False, True, True, True], - [False, False, False, False, False], - ], - ] - ), - 0.5, - np.array([True, True]), - DoesNotRaise(), - ), # two masks partially overlapping with different category - ( - np.array( - [ - [0, 0, 0, 0, 0.8], - [0, 0, 0, 0, 0.85], - [0, 0, 0, 0, 0.9], - ] - ), - np.array( - [ - [ - [False, False, False, False, False], - [False, True, True, False, False], - [False, True, True, False, False], - [False, False, False, False, False], - [False, False, False, False, False], - ], - [ - [False, False, False, False, False], - [False, True, True, False, False], - [False, True, True, False, False], - [False, False, False, False, False], - [False, False, False, False, False], - ], - [ - [False, False, False, False, False], - [False, False, False, True, True], - [False, False, False, True, True], - [False, False, False, False, False], - [False, False, False, False, False], - ], - ] - ), - 0.5, - np.array([False, True, True]), - DoesNotRaise(), - ), # three masks with no category - ( - np.array( - [ - [0, 0, 0, 0, 0.8, 0], - [0, 0, 0, 0, 0.85, 1], - [0, 0, 0, 0, 0.9, 2], - ] - ), - np.array( - [ - [ - [False, False, False, False, False], - [False, True, True, False, False], - [False, True, True, False, False], - [False, False, False, False, False], - [False, False, False, False, False], - ], - [ - [False, False, False, False, False], - [False, True, True, False, False], - [False, True, True, False, False], - [False, True, True, False, False], - [False, False, False, False, False], - ], - [ - [False, False, False, False, False], - [False, True, True, False, False], - [False, True, True, False, False], - [False, False, False, False, False], - [False, False, False, False, False], - ], - ] - ), - 0.5, - np.array([True, True, True]), - DoesNotRaise(), - ), # three masks with different category - ], -) -def test_mask_non_max_suppression( - predictions: np.ndarray, - masks: np.ndarray, - iou_threshold: float, - expected_result: Optional[np.ndarray], - exception: Exception, -) -> None: - with exception: - result = mask_non_max_suppression( - predictions=predictions, masks=masks, iou_threshold=iou_threshold - ) - assert np.array_equal(result, expected_result) - - @pytest.mark.parametrize( "xyxy, resolution_wh, expected_result", [ @@ -1033,8 +720,8 @@ def test_calculate_masks_centroids( ), # two data dicts with the same field name and different length arrays values ( [{}, {"test_1": [1, 2, 3]}], - {"test_1": [1, 2, 3]}, - DoesNotRaise(), + None, + pytest.raises(ValueError), ), # two data dicts; one empty and one non-empty dict ( [{"test_1": [], "test_2": []}, {"test_1": [1, 2, 3], "test_2": [1, 2, 3]}], diff --git a/test/utils/test_internal.py b/test/utils/test_internal.py new file mode 100644 index 00000000..eee614e6 --- /dev/null +++ b/test/utils/test_internal.py @@ -0,0 +1,193 @@ +from contextlib import ExitStack as DoesNotRaise +from dataclasses import dataclass, field +from typing import Any, Set + +import numpy as np +import pytest + +from supervision.detection.core import Detections +from supervision.utils.internal import get_instance_variables + + +class MockClass: + def __init__(self): + self.public = 0 + self._protected = 1 + self.__private = 2 + + def public_method(self): + pass + + def _protected_method(self): + pass + + def __private_method(self): + pass + + @property + def public_property(self): + return 0 + + @property + def _protected_property(self): + return 1 + + @property + def __private_property(self): + return 2 + + +@dataclass +class MockDataclass: + public: int = 0 + _protected: int = 1 + __private: int = 2 + + public_field: int = field(default=0) + _protected_field: int = field(default=1) + __private_field: int = field(default=2) + + public_field_with_factory: dict = field(default_factory=dict) + _protected_field_with_factory: dict = field(default_factory=dict) + __private_field_with_factory: dict = field(default_factory=dict) + + def public_method(self): + pass + + def _protected_method(self): + pass + + def __private_method(self): + pass + + @property + def public_property(self): + return 0 + + @property + def _protected_property(self): + return 1 + + @property + def __private_property(self): + return 2 + + +@pytest.mark.parametrize( + "input_instance, include_properties, expected, exception", + [ + ( + MockClass, + False, + None, + pytest.raises(ValueError), + ), + ( + MockClass(), + False, + {"public"}, + DoesNotRaise(), + ), + ( + MockClass(), + True, + {"public", "public_property"}, + DoesNotRaise(), + ), + ( + MockDataclass(), + False, + {"public", "public_field", "public_field_with_factory"}, + DoesNotRaise(), + ), + ( + MockDataclass(), + True, + {"public", "public_field", "public_field_with_factory", "public_property"}, + DoesNotRaise(), + ), + ( + Detections, + False, + None, + pytest.raises(ValueError), + ), + ( + Detections, + True, + None, + pytest.raises(ValueError), + ), + ( + Detections.empty(), + False, + {"xyxy", "class_id", "confidence", "mask", "tracker_id", "data"}, + DoesNotRaise(), + ), + ( + Detections.empty(), + True, + { + "xyxy", + "class_id", + "confidence", + "mask", + "tracker_id", + "data", + "area", + "box_area", + }, + DoesNotRaise(), + ), + ( + Detections(xyxy=np.array([[1, 2, 3, 4]])), + False, + { + "xyxy", + "class_id", + "confidence", + "mask", + "tracker_id", + "data", + }, + DoesNotRaise(), + ), + ( + Detections( + xyxy=np.array([[1, 2, 3, 4], [5, 6, 7, 8]]), + class_id=np.array([1, 2]), + confidence=np.array([0.1, 0.2]), + mask=np.array([[[1]], [[2]]]), + tracker_id=np.array([1, 2]), + data={"key_1": [1, 2], "key_2": [3, 4]}, + ), + False, + { + "xyxy", + "class_id", + "confidence", + "mask", + "tracker_id", + "data", + }, + DoesNotRaise(), + ), + ( + Detections.empty(), + False, + {"xyxy", "class_id", "confidence", "mask", "tracker_id", "data"}, + DoesNotRaise(), + ), + ], +) +def test_get_instance_variables( + input_instance: Any, + include_properties: bool, + expected: Set[str], + exception: Exception, +) -> None: + with exception: + result = get_instance_variables( + input_instance, include_properties=include_properties + ) + assert result == expected