Sport Video Analysis for Billiard Matches

This project aims to develop a computer vision system for analyzing video footage of "Eight Ball" billiard games. The system will leverage computer vision techniques to provide real-time insights into the game's state, including ball positions, trajectories, and game progress. The output will be visualized as a 2D top-view minimap of the playing field.

Authors

• @Mohammadhossein Akbari
• @Zahra_Rahgooy
• @Giacomo_D’Andria

Features

• Playing Field Detection: Identifies the boundaries of the playing field.

• Ball Detection, Localization and Segmentation: Recognizes and tracks all balls on the table and segments them into categories.

• Top-View Visualization: Updates a 2D minimap with current ball positions and trajectories.

Instructions

$ mkdir build

$ cd build

$ cmake -DCMAKE_BUILD_TYPE=Release ..

$ make

$ cd .. && cd bin

Usage Example:

Usage: ./Starter < Input video path > < Output video path >