This project is part of the Signal, Image, and Video course taught by Prof. Andrea Rosani for the academic year 2023/2024. As an introduction course to Computer Vision, my goal was to apply traditional methods like Histogram of Oriented Gradients, CAMShift and Kalman filters, to identify and track with colored (with the team color) bounding boxes all football players in a selected video.

It is far from perfect, as you can see in the final result below, but I really enjoyed taking the first step into the field of CV applied to sport.

If you want to understand my development process and the problems I encountered, this is the report . Feel free to write me an email if you want to discuss some points or suggest upgrades.

Ensure you have Python installed on your system. This project requires the use of a virtual environment to manage dependencies.

1. Clone the repository to your local machine.
2. Navigate to the cloned directory and create a virtual environment:

python -m venv venv

3. Activate the virtual environment:

• On Windows:

  venv\Scripts\activate
  

• On Unix or MacOS:

  source venv/bin/activate
  

4. Install the required libraries:

pip install -r requirements.txt

Execute the main script to start the player detection and tracking process:

python src/main.py

The results will be saved in the Results directory.

• /Data
  • /bounding-boxes/frame.json: Contains the JSON with detected coordinates corresponding to the video in results. This file will be overwritten with new detections upon subsequent runs.
  • /video/Bundes clip.mp4: The input video file.
• /Results
  • /video/player-detection.avi: The final result of the analysis.
• /src
  • drawVideo.py: Functions for drawing bounding boxes on the video.
  • function.py: Common utility functions.
  • HOG.py: Functions for detecting new players with HOG.
  • KALMShift.py: Functions for tracking bounding boxes in memory.
  • main.py: The main script to run the analysis.
  • matchBoxes.py: Functions for matching boxes from HOG detection and the tracker.
  • utilities.py: File containing variable declarations, kernels, and paths.
• requirements.txt: Libraries required for the virtual environment.
• README.md: This document.