/RefereeChallenge

UvA bsc Artificial Intelligence project (2022)

Primary LanguageJupyter NotebookCreative Commons Attribution 4.0 InternationalCC-BY-4.0

Referee Challenge

UvA batchelor thesis project, which aims to resolve "The Visual Referee Challenge". This solution to the challenge aims to combine whistle detection and pose estimation, in order to derive referee rulings.

Layout

Whistle detection

in the whistle_detection folder, open whistle_detection.ipynb and run the notebook. The dataset is already in the directory, but can also be downloaded from the original source.

Whistle Dataset

Pose Estimation

in the pose_estimation folder, open pose_estimation_basic.ipynb or pose_estimation_advanced.ipynb and run the notebook. The dataset needs to be downloaded seperately (too large for github). Put the dataset in the same directory as the notebooks.

Referee Dataset

Thesis

The PDF of the thesis can be found at the following link:

Thesis