🎨 Digit Recognition Drawing App 🖌️

Welcome to the Digit Recognition Drawing App! This project allows users to draw digits on a 28x28 canvas, and the AI model predicts the digit using a neural network (MLP Classifier). The app is built using Tkinter for the GUI and scikit-learn for the machine learning model.

---

✨ Features

• 🎨 Interactive Drawing Canvas: Draw digits freely on a 28x28 grid.
• 🤖 AI-powered Prediction: Uses a neural network (MLP) for better accuracy.
• 🧹 Clean Board: Reset the canvas with a single click.
• 📈 Pre-trained Model Loading: Avoid re-training with a pre-saved model.

---

🛠️ Prerequisites

• Python 3.x
• Required libraries:
  • numpy
  • matplotlib
  • scikit-learn
  • tkinter (Usually pre-installed with Python)
  • joblib

Install the required libraries using:

pip install numpy matplotlib scikit-learn joblib

---

📦 Files Included

• main.py - The main Python file to run the app.
• mnist_mlp_model.pkl - Pre-trained MLP model (auto-generated after training).

---

🚀 How to Run the App

1. Clone the Repository:

   git clone https://github.com/keshav-kh/digit-recognition-app.git

2. Navigate to the Project Directory:

   cd digit-recognition-app

3. Run the Application:

   python main.py

After the application starts, draw a digit on the canvas, and click "Predict Digit". The AI will predict the digit and display it. Use "Clean Board" to reset the canvas.

---

📊 Model Training

The application uses the MNIST dataset and an MLP Classifier from scikit-learn. The model is trained once and saved as mnist_mlp_model.pkl to skip re-training every time the app is run.

If you wish to re-train the model, delete the .pkl file, and the app will automatically train a new model.

---

🤝 Contribution

Feel free to open issues or submit pull requests. Any improvements to the model or the app's user interface are welcome!

---

📝 License

This project is open-source and available under the MIT License.

---

Have fun drawing! 🖌️✨