Computer vision (CV) has shown great promise as of late within the deep learning ecosystem. There are many applications for which CV is being used. However, one that has not been widely explored yet is the use of CV in sports to automatically detect when a foul occurs. Automating this process could by in large help perhaps eliminate issues that arise when the outcome of a game is left to the referee's discretion, which can be controversial at times. Here I attempt to leverage the power of CV to make strides towards automating the process of calling fouls in basketball.
If you would like a comprehensive description of the project, listen to the short video at demo_videos/presentation_video.mp4!
To view the end-to-end project solution, please checkout the Jupyter Notebook by clicking on automated_and1.ipynb.
Slides are available at presentation.pdf.