Handwritten Digit Recognition using CNN

This project is a Handwritten Digit Recognition system using Convolutional Neural Networks (CNNs). It is implemented using TensorFlow and Keras.

Features

• Trains a CNN model on the MNIST dataset.
• Evaluates the model's accuracy on test data.
• Visualizes training history using Matplotlib.
• Displays sample predictions from the test dataset.

Installation

1. Clone the Repository

git clone https://github.com/yourusername/handwritten-digit-recognition.git
cd handwritten-digit-recognition

2. Install Dependencies

Ensure you have Python installed, then install the required libraries:

pip install tensorflow numpy matplotlib

Usage

Train and Test the Model

Run the following command in a Jupyter Notebook script:

python handwritten_recognition.ipynb

This will:

• Load and preprocess the MNIST dataset.
• Train a CNN model.
• Evaluate its accuracy.
• Display sample predictions.

Model Architecture

• Conv2D (32 filters, 3x3, ReLU activation)
• MaxPooling2D (2x2)
• Conv2D (64 filters, 3x3, ReLU activation)
• MaxPooling2D (2x2)
• Conv2D (64 filters, 3x3, ReLU activation)
• Flatten Layer
• Dense (64 neurons, ReLU activation)
• Dense (10 neurons, Softmax activation)

Results

The model achieves an accuracy of 98%+ on the MNIST test set.

Screenshots

License

This project is licensed under the MIT License.

Contributions

Feel free to fork the repository and submit a pull request if you want to improve the model or add new features.

Contact

For any questions, open an issue in the repo.