Merge pull request #1259 from Vishisht16/main

Sudoku Solver feature

Author: sanjay-kv

OpenCV Projects/Sudoku Solver/README.md

@@ -0,0 +1,28 @@
+# Sudoku Solver
+
+This project is a Sudoku solver that reads a Sudoku puzzle from an image, recognizes the digits using a pre-trained OCR model, and then solves the puzzle. The final solved puzzle is displayed by overlaying the solution onto the original image.
+
+## Features
+- Extracts the Sudoku grid from an image using image processing techniques.
+- Recognizes the digits in the puzzle using a pre-trained OCR model.
+- Solves the puzzle using a backtracking algorithm.
+- Overlays the solution onto the original image.
+
+## How it Works
+
+### 1. Image Preprocessing:
+   The input image is processed to detect the Sudoku grid. Contours are used to locate the largest rectangular region in the image, which is assumed to be the Sudoku puzzle.
+
+### 2. Digit Recognition:
+   The extracted grid is split into 81 cells, and each cell is resized to match the input size of the pre-trained OCR model. The model predicts the digits in the puzzle, assigning a number to each cell or marking it as empty.
+
+### 3. Puzzle Solving:
+   The recognized digits are passed into a backtracking algorithm, which solves the Sudoku puzzle. If the puzzle cannot be solved, the model might have misread a digit.
+
+### 4. Output:
+   The solved digits are overlaid back onto the original image using inverse perspective transformation, showing the solution in the correct position.
+
+## Usage
+
+1. Place your Sudoku puzzle image in the project directory (e.g., `sudoku1.jpg`).
+2. Run the script.