An AI-powered web application that detects handwritten digits from an uploaded A4 sheet photo using a Convolutional Neural Network (CNN) model built with TensorFlow and deployed using Streamlit.
This project allows users to:
- Upload a photo of an A4 sheet containing a handwritten digit inside a bold square.
- Automatically crop, preprocess, and convert the image to a 28×28 grayscale format.
- Predict which number (0–9) was written using a trained CNN model.
- Display the prediction and confidence score.
- 🐍 Python 3.10+
- 🧠 TensorFlow / Keras
- 🖼️ OpenCV (image processing)
- 🌐 Streamlit (web app UI)
- 📦 NumPy, Pillow