Handwritten Digit Classification with Convolutional Neural Networks

This project implements a Convolutional Neural Network (CNN) to classify handwritten digits (0 to 9) using the popular MNIST dataset. The CNN architecture is built using TensorFlow/Keras and achieves accurate digit recognition through a series of convolutional and pooling layers.

Features

• Data Loading: Utilizes Pandas to load handwritten digit data from CSV files.
• Data Preprocessing: Reshapes and normalizes the data, preparing it for CNN input.
• Convolutional Neural Network (CNN): Implements a CNN model using TensorFlow/Keras layers.
• Model Training: Trains the CNN model on the MNIST dataset for handwritten digit recognition.
• Model Evaluation: Evaluates the trained model's accuracy on a validation dataset.
• Prediction: Provides a function to make predictions on new handwritten digit images.

Requirements

• Python 3.x
• Pandas
• TensorFlow/Keras
• scikit-learn

Project Structure

• main.py: Main Python file containing the code for data loading, preprocessing, model creation, training, evaluation, and prediction.
• data/: Directory containing the CSV files with the training and testing data.
• models/: Directory to store saved trained models.

Acknowledgements

• MNIST Dataset: The MNIST dataset used in this project is publicly available and widely used for handwritten digit recognition tasks. More information can be found here.