This is an image-based sudoku solver project that takes an image as input, and then uses OpenCV to find out the Sudoku and transform it into a 2D vector, and then solve it and display a solved sudoku. The input images are screenshots of the games from the Sudoku game on Play Store. This project is developed on Docker container for better portability and application isolation.
This project is just a simple project with some contraints/assumptions such as: 1) Input images are screenshots from the 'Sudoku' game available on PlayStore 2) The font style of the digits are the same 3) The sudoku is the biggest contour on the image
Based on the constraints/assumptions above, using OpenCV template matching is enough to perform digit recognition.
./buildDockerImage.sh
./runDockerImage.sh