Creating a system that track the game and show the movements and records at the end of the match.
Only image processing techniques were used to create the system.
There isn't any deep learning model or machine learning algorithm.
We can simply list the processing steps as below
- Convert the board image into Bird'eye view
- Check if hand is on the board
- Find checker contours
- Find the position of each checker on the board
- Seperate white and black checkers
- Compare previously saved state with the current state and find the differences
- Save the changes (Player, Checkter Old Position, Checker New Position, Defeated Rival Checkers)
Original Video Frame:
Bird'eye View Frame:
When the player makes a move
When player defeat a rival checker
