🧠 Handwritten Digit Recognition with TensorFlow/Keras

A beginner-friendly AI project to classify handwritten digits (0–9) using a Convolutional Neural Network (CNN) on the MNIST dataset. Built with TensorFlow/Keras.

---

📌 Features

• CNN model for image classification
• 98–99% accuracy on test data
• Visualizes predictions and training metrics
• Includes model saving/loading

---

📁 Dataset

• MNIST Handwritten Digits

• 60,000 training images, 10,000 test images

• Grayscale, 28x28 pixels

---

🧰 Tech Stack

• Python
• TensorFlow & Keras
• NumPy, Matplotlib
• Jupyter Notebook

---

🚀 Getting Started

1. Clone the Repository

git clone https://github.com/lina2016/mnist-digit-recognition.git
cd mnist-digit-recognition



2. Install Dependencies

pip install tensorflow matplotlib numpy



3. Run the Notebook

Launch Jupyter Notebook:
jupyter notebook mnist_digit_recognition.ipynb



📈 Results

Test Accuracy: ~98.5%
Example Prediction:



💡 Future Improvements

Add data augmentation
Try different architectures or optimizers
Deploy with Streamlit or Flask
Expand to Fashion MNIST or custom digit images



📜 License

This project is licensed under the MIT License.

🙋‍♂️ Author

This project was created by Lina Jamadar as a personal learning exercise using the MNIST dataset and TensorFlow/Keras.
Inspired by various public tutorials and documentation.