Sudoku Image Solver

Solves sudoku given as an image using OpenCV Python and Python Tessaract OCR

Roughly summarized steps

1. Preprocess the image includes sudoku board (hint: use perspective transform and make it binary)
2. Divide the board into 9*9=81 cells
3. Extract numbers from each cell (OCR) and store them
4. Solve the sudoku using backtracking algorithm
5. Show the result on the given sudoku board (as you can see below)

Things I can improve

Considering a better way to preprocess the given image.
Implementing deep learning method to extract numbers.

Result