Employing Image Processing to Extract the Sudoku Puzzle from an Image and solving it with a Backtracking Algorithm.

Using the Backtracking Algorithm Solve the unsolved Sudoku. The approach for solving Sudoku puzzles comprises extracting the puzzle from a picture using image processing techniques, and then solving the puzzle using a backtracking algorithm. To detect the circles and digits in the image, the initial part of the procedure entails pre-processing the image using a variety of methods, including adaptive thresholding and the Hough Transform. The circles and digits can be separated from the rest of the image by using these techniques to identify their contour. A pre-trained model is used to recognize the numbers and store them in the appropriate places in the puzzle after the circles and digits have been detected. This model can properly identify the digits even when the image quality is poor or the problem has been captured at an angle because it was trained on a big dataset of photographs of Sudoku puzzles. The method uses a backtracking algorithm to solve the puzzle after the recognized digits have been added to it. This method operates by repeatedly attempting various digit combinations in the puzzle's empty cells, and iterating backwards anytime it comes upon a contradiction. Overall, the technique can extract the puzzle from the image and correctly identify the numbers, enabling it to finish the puzzle rapidly and show the solution. [ABSTRACT FROM AUTHOR]