✍️ Handwritten Math Solver

A Flask application that uses OCR + CNN to recognize handwritten digits and math operators from an image, reconstruct the expression, and safely evaluate it (no insecure eval()).

---

📸 Example

➡️ The app predicts 2+3*4 and computes the result: = 14.

---

🔧 Features

• 🖼️ Image preprocessing with OpenCV
• 🔤 CNN model (Keras/TensorFlow) for digit/operator recognition
• 🧮 Safe arithmetic evaluator (no arbitrary code execution)
• 🌐 Flask API endpoint /predict for uploads
• ✅ Unit tests and GitHub Actions CI

---

⚡ Quick Start

Option 1: Using Docker (Recommended)

Prerequisites

• Docker installed on your system

Steps

1. Clone the repository

   git clone https://github.com/iot-lab-kiit/handwritten-math-solver.git
   cd handwritten-math-solver

2. Add model files
   Place your trained model files in the model/ folder:

   • model/model.json
   • model/model_weights.h5

3. Build the Docker image

   docker build -t handwritten-math-solver .

4. Run the container

   docker run -p 5000:5000 handwritten-math-solver

5. Access the app
   Visit http://localhost:5000

---

Option 2: Local Setup

1. Clone & install

git clone https://github.com/iot-lab-kiit/handwritten-math-solver.git
cd handwritten-math-solver
python -m venv venv
source venv/bin/activate   # or venv\Scripts\activate on Windows
pip install -r requirements.txt

2. Add model files

Place your trained model files in the model/ folder:

• model/model.json
• model/model_weights.h5

3. Run the server

python app.py

Visit http://127.0.0.1:5000.

---

🤝 Contributing

We welcome contributions!
Check issues labeled “good first issue” or “help wanted” if you’re new.

Ways to help:

• Improve preprocessing or recognition
• Add more operators or equation solving
• Write more tests
• Improve docs or CI/CD pipeline

See contributing.md for details.

---

📚 Tech Stack

• Python 3.9+
• Flask
• TensorFlow / Keras
• OpenCV
• NumPy / Pillow
• Pytest
• GitHub Actions

---

📄 License

MIT License — see LICENSE