Handwritten-Digit-Recognition-using-CNN

Overview

This project uses Convolutional Neural Networks (CNN) to recognize handwritten digits. Trained on the MNIST dataset, the model can accurately predict single and double-digit numbers from user input or uploaded images. A graphical user interface (GUI) built with Tkinter provides an interactive way to draw or upload images for model predictions.

Features

• Handwritten Digit Recognition: The model identifies digits using a CNN-based architecture trained on the MNIST dataset.
• User Interface: A GUI application built with Tkinter allows users to draw digits and see predictions in real-time.
• Model Persistence: The model can be saved and reloaded for further use without retraining.

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/Handwritten-Digit-Recognition-CNN.git

2. Navigate into the project directory:

   cd Handwritten-Digit-Recognition-CNN

3. Install the required packages:

   pip install -r requirements.txt

Usage

1. Run the Jupyter Notebook to train and evaluate the CNN model on the MNIST dataset.
2. Launch the GUI using the provided Tkinter application script to interact with the model:

   python gui_application.py

3. Draw or upload an image of a handwritten digit to see the model’s prediction.

Model Training

The CNN model is trained on the MNIST dataset using TensorFlow and Keras. You can adjust model parameters and training settings in the provided Jupyter Notebook.

Acknowledgments

• The MNIST dataset for digit images.
• TensorFlow and Keras for model building.
• Tkinter for the GUI.