DeepSport Utilities toolkit
This toolkit offers a wide range of helpers to download, transform and use datasets following the DeepSport format.
Installation
Package is released on PyPI for convenient installation:
pip install deepsport-utilities
For developement, install with:
git clone
pip install -e ".[dev]"
Available datasets
- Basketball Instants Dataset is implemented by
deepsport_utilities.InstantsDatasetand provides raw images captured by the Keemotion system at different instants. - Basketball Ballistic Trajectories Dataset, implemented as multiple successive instants.
- Private Keemotion datasets.
Dataset format
The datasets are stored as a json file holding items metadata and multiple data files. The easiest approach to load the data is to use the import_dataset function as illustrated in the scripts provided in the examples folder.
The resulting datasets are based on mlworkflow.Dataset, a dataset implementation of (key, value) pairs where the keys are light and allow efficient querying of the dataset while the values contain the heavy data. For more information, refer to mlworkflow repository.
Toolkit
Along with the provided datasets, the library comes with utility functions to process the dataset work with basketball courts.
Working with calibrated images captured by the Keemotion system.
Calibration data are implemented with cailb3d.Calib objects, allowing determining the relation between the image pixels (2D coordinates) and points in the real world (3D coordinates).
Images shared by Keemotion follow a convention where the origin of the 3D world is located on the furthest left corner of the basketball court relative to the main camera setup ; more precisely in the inner side of the court lines. The unit of length is the centimeter and axis orientation is illustrated here, with z pointing downward:
Contributing
This library is open-source and contributions are welcome. However, prior to any implementation, a discussion with the main maintainer Gabriel Van Zandycke is required.
Authors and acknowledgment
While most of the library was developed by Gabriel Van Zandycke, this library benefited from the work of
- Maxime Istasse for the project initial kick-off
- Cedric Verleysen for some functions in
court.py
1 maintainer
Metadata
1 maintainer
Release files for deepsport-utilities 5.2.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| deepsport_utilities-5.2.1.tar.gz | 2.9 MB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| deepsport_utilities-5.2.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 2.9 MB
Release files / deepsport_utilities-5.2.1.tar.gz
| Download URL | deepsport_utilities-5.2.1.tar.gz |
|---|---|
| Size | 2.9 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
85aab78f91a5ad262860e1b222e6bd8aead046685b92129ceef98da09278e718
|
|
BLAKE2b-256 checksum How to use checksums |
a9ee634629b0f926d177a0619dd2f7bad4cda64c17a843667002b20cd905be6a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.1.0 CPython/3.10.12
|
Release files / deepsport_utilities-5.2.1-py3-none-any.whl
| Download URL | deepsport_utilities-5.2.1-py3-none-any.whl |
|---|---|
| Size | 44.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
b95917db1f12a335de59c91d7e26814d3f991ba3311e358e737b0d191e488bf6
|
|
BLAKE2b-256 checksum How to use checksums |
7ab240fcc6eb27356ddee30bd1ace9feb184460c6b2a36d0a4416930df16ce59
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.1.0 CPython/3.10.12
|
