Welcome to your step-by-step guide on creating an AI Sudoku Solver that can extract puzzles from photos and solve them! This project combines computer vision with artificial intelligence to give you a creative way to tackle Sudoku puzzles, whether you’re a beginner or a seasoned pro.

Table Of Contents:

• Installation
• Usage
• Working
  • Image Preprocessing
  • Recognition
• ToDo
• Contributing

Installation

To set up your AI Sudoku Solver, follow these steps:

1. Download and install Python3 from here.
2. Install virtualenv by typing the following command in a terminal:

pip install virtualenv

3. Create a virtual environment named sudokuenv:
   For Windows:

   virtualenv sudokuenv
   cd sudokuenv\Scripts
   activate

   For Linux:

   source sudokuenv/bin/activate

4. Clone the repository and navigate into the cloned folder:

cd A:AI_Sudoku-master

5. Install the required packages:

pip install -r requirements.txt

Usage

Before running the application, you’ll need to set the modeltype variable in Run.py. The available options are:

• CNN for the Convolutional Neural Network
• KNN for K Nearest Neighbours Algorithm (default)

To run the application, execute the following command:

python Run.py

You need an internet connection when creating the knn.sav file, which might take 5-10 minutes during the first run.

Working

Now let’s explore the inner workings of the AI Sudoku Solver, particularly image preprocessing and recognition. We’ll clarify how it processes the Sudoku image using an analogy of baking a cake!

Image Preprocessing

Imagine you’re making a layered cake (the Sudoku board) and need to gather all your ingredients (image data) first. Here’s how image preprocessing can be compared to your baking prep, focusing on how each step enhances your cake’s outcome:

• Gaussian Blurring: This step is like sifting flour; it reduces noise in the image, ensuring a fluffier outcome.
• Adaptive Gaussian Thresholding: You adjust your oven’s temperature to ensure even cooking—this filters out light variations in the image.
• Inverting: Just as you would grease your cake pans to prepare for baking, this flips colors so the digits are white and the background black.
• Dilation: This adds a rich frosting layer, thickening lines for better structure.
• Flood Filling: Much like pouring batter into the pan, it fills in areas, helping to identify the largest ‘component’—the Sudoku grid.
• Eroding: This is the visual touch-up before the final bake, removing excess from the outer layer.
• Hough Line Transform: Picture this as assembling your cake layers thoughtfully — it identifies where each cake slice is by finding straight edges.
• Warping Perspective: Finally, like placing the final layer of icing smoothly, it corrects the perspective for clarity.

Recognition

Next, we have the actual baking or recognition phase, where the prepared ingredients (image data) are baked into a delicious cake (the final Sudoku solution). Here, the cake can be baked with:

• Convolutional Neural Network (CNN): Think of this as choosing a carefully curated recipe that ensures a delicious result.
• K Nearest Neighbours (KNN): This is like a family recipe shared among friends, tested and proven to be reliable.

ToDo

• Improve accuracy.
• Resolve any bugs or issues.
• Optimize code for better performance.

Contributing

Your contributions are appreciated! To keep the project lively, follow these guidelines for pulling requests:

• Create clean code with comments.
• Test the application thoroughly.
• Open a pull request; your efforts will be acknowledged!

Troubleshooting

If you encounter any issues, here are some tips to resolve them:

• Ensure that all packages from requirements.txt are installed correctly.
• Verify that your Python environment is activated appropriately.
• If the program crashes, check for any logs that might indicate missing libraries or improper configurations.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.