In this repository there are 3 implementations:
- Android-Java: Basic implementation including image capture and game management. The Android app game management is not finished, but the computer vision and OCR part is working.
- C++: Basic OCR prototype. Gets an input image and retrieves the letters on the board.
- Python: Basic OCR prototype. Gets an input image and retrieves the letters on the board.
- Detect the outer frame of the board according to its red color, using HSV color coordinates.
- Find the 4 corners of the board and arrange them by order.
- Apply perspective transformation to compensate the camera view angle.
- Detect each tile square and decide if it is empty or have a letter on it, by looking on the STD or gradients strength.
- For each letter we found, we call the Tesseract optical character recognition.
Android app:
