Air Reader : Gesture-Controlled PDF Viewer

Hands-free PDF navigation using:

• Hand gesture recognition (MediaPipe)
• KNN classifier (98.2% accuracy)
• PDF.js integration

Features

• Zoom in/out with pinch gestures
• Real-time hand tracking
• Cross-platform (Windows/macOS/Linux)

Installation

git clone https://github.com/ajydeep/Air-Reader.git
cd Air-Reader
pip install -r requirements.txt

Usage

1. Run data collection:

   python collect_data.py

2. Train the model:

   python train_model.py

3. Start the viewer:

   python main.py

Project Structure

├── pre/                  
│   ├── collect_data.py
│   ├── best_gesture_model.pkl
│   └── train_model.py
├── data/                 
│   └── gesture_data.csv
├── main.py          
├── viewer.html
└── README.md

License

MIT