Handwritten Digit Recognition using CNN

Overview

This project implements a Handwritten Digit Recognition system using a Convolutional Neural Network (CNN). The system consists of two main parts:

1. Model Training (program.py) - A CNN model is trained on the MNIST dataset.
2. Digit Recognition GUI (practiceground.py) - A graphical interface allows users to draw digits, which are then classified using the trained model.

---

Features

• CNN-based Model for digit classification.
• Tkinter GUI for drawing and recognizing digits.
• Uses OpenCV for image processing.
• Model is saved in mnist.h5 for later use.

---

Technologies Used

• Python
• TensorFlow/Keras
• OpenCV
• Tkinter
• NumPy

---

Dataset Used

• MNIST Dataset: A collection of 60,000 training and 10,000 testing images of handwritten digits (0-9).

---

File Descriptions

1. Model Training (program.py)

This script trains a CNN on the MNIST dataset and saves the model (mnist.h5).

Key Steps:

• Loads MNIST dataset.
• Preprocesses the images (resizing, normalizing, reshaping).
• Builds a CNN with:
  • Convolutional layers for feature extraction.
  • MaxPooling layers for dimensionality reduction.
  • Dense layers for classification.
• Trains the model and evaluates performance.
• Saves the trained model for later use.

Run the Training Script:

python train_model.py

---

2. Handwritten Digit Recognition GUI (practiceground.py)

This script provides a graphical interface where users can draw a digit, and the model predicts the digit.

Key Features:

• Tkinter Canvas: Users can draw digits.
• Clear Button: Reset the canvas.
• Predict Button: Processes the image and predicts the digit.
• Uses OpenCV to process the drawn image before passing it to the model.

Run the GUI Application:

python main.py

---

How It Works

1. Train the model using program.py. This saves the trained model as mnist.h5.
2. Run the GUI using practiceground.py.
3. Draw a digit in the Tkinter window.
4. Click the "Predict" button to classify the digit.
5. The predicted digit is displayed in a new window.

---

Requirements

Install dependencies using:

pip install numpy opencv-python tensorflow keras

---

Future Enhancements

• Improve model accuracy with more training.
• Add support for different handwriting styles.
• Deploy as a web-based application.

---

Credits

• Dataset: MNIST (by Yann LeCun)
• Libraries: TensorFlow, OpenCV, Tkinter