Installation and Setup Instructions

""" INSTALLATION GUIDE:

1. Create a new Python virtual environment: python -m venv auth_system_env

2. Activate the virtual environment:

   • Windows: auth_system_env\Scripts\activate
   • Linux/Mac: source auth_system_env/bin/activate

3. Install required packages: pip install -r requirements.txt

4. Additional setup required:

   a) For dlib (face landmark detection):

   • Download shape_predictor_68_face_landmarks.dat from: http://dlib.net/files/shape_predictor_68_face_landmarks.dat.bz2
   • Extract and place in data/models/ directory

   b) For PyAudio (voice recording):

   • Windows: pip install pipwin && pipwin install pyaudio
   • Linux: sudo apt-get install portaudio19-dev && pip install pyaudio
   • Mac: brew install portaudio && pip install pyaudio

5. Create project directory structure: mkdir -p data/models data/users config models utils tests

6. Run the system: python main.py

VSCODE SETUP:

1. Install Python extension for VSCode
2. Select Python interpreter: Ctrl+Shift+P -> "Python: Select Interpreter"
3. Choose the virtual environment interpreter
4. Install additional VSCode extensions:
   • Python Docstring Generator
   • Python Test Explorer
   • Jupyter (for model training notebooks)

USAGE:

1. Run main.py to start the authentication system
2. Choose option 1 to register a new user
3. Choose option 2 to authenticate an existing user
4. Follow on-screen instructions for face and voice authentication

FEATURES:

✅ Multi-modal authentication (Face + Voice + VSR) ✅ Advanced liveness detection ✅ Real-time video processing ✅ 3D-CNN based Visual Speech Recognition ✅ BiGRU-BiLSTM fusion architecture ✅ Multi-head attention mechanism ✅ Comprehensive preprocessing pipeline ✅ Robust face detection and recognition ✅ Anti-spoofing measures ✅ Modular and extensible design ✅ Unit testing framework

PERFORMANCE NOTES:

• The system requires a decent GPU for real-time VSR processing
• CPU-only mode will work but may be slower
• Minimum 8GB RAM recommended
• Camera with at least 720p resolution recommended

CUSTOMIZATION:

• Adjust thresholds in config/settings.py
• Modify model architectures in models/ directory
• Add new authentication methods by extending base classes
• Customize preprocessing in utils/ directory """