🔢 MNIST Number Prediction

Predict handwritten digits using a trained deep learning model and an interactive drawing canvas built with Tkinter.

---

📌 Project Overview

This project demonstrates a simple yet effective way to train a digit recognition model on the MNIST dataset and use it in a graphical interface for real-time predictions.

• 🎯 Train your own model with model.py
• 🎨 Draw digits and predict them live with drawingPredict.py

---

✨ Features

• 🧠 Trains a deep learning model using TensorFlow/Keras
• 🖌️ Interactive canvas using Tkinter
• 💾 Saves and loads trained model automatically
• 🔍 Real-time digit recognition after drawing

---

🛠️ Tech Stack

Tool	Purpose
Python	Programming Language
TensorFlow/Keras	Model Training
Tkinter	GUI Drawing Canvas
NumPy & PIL	Image Processing
MNIST Dataset	Training Dataset

---

🚀 Getting Started

1️⃣ Clone the Repository

git clone https://github.com/vaishnavdounde65/MNIST-Number-Prediction.git
cd mnist-number-predictor

2️⃣ Install Dependencies

pip install PIL numpy tensorflow tkinter

Note: Make sure Python ≥ 3.6 is installed.

🧠 Train the Model

Run the script to train and save the model:

python model.py

Trains on the MNIST dataset

Saves the model as mnist_model.h5

🖼️ Predict with Drawing Interface

Use the drawing interface to draw digits and predict:

python drawingPredict.py

A Tkinter window will open

Draw a digit with your mouse

Click Predict to see the model’s guess!

📁 File Structure

├── model.py              # Train and save the model
├── drawingPredict.py     # GUI for drawing and predicting
├── mnist_model.h5        # Saved model (after training)
└── README.md             # Project overview

📸 Demo

🙌 Acknowledgements

MNIST Dataset - Yann LeCun

TensorFlow/Keras

Tkinter GUI Library

📬 Contact

Vaishnav Dounde 🌐 vaishnav.site

⭐ Star This Repository

If you found this project useful, feel free to ⭐ the repo and share it!