This project is a real-time face recognition system that can identify known individuals from a live video stream. It uses the face_recognition library, which is built on top of dlib.

• Encodes faces from a directory of images.
• Performs real-time face recognition using a webcam.
• Displays the name of the recognized person on the video stream.

live_face_detection
├── .flake8
├── .gitignore
├── run.sh                  # Main script to run the project
├── requirements.txt        # Python dependencies
├── source_images/          # Directory for source images of people to recognize
└── src/
    ├── encode_faces.py     # Script to encode faces from source_images
    ├── live_recognition.py # Script to run live face recognition
    ├── face_encodings/
    │   └── encodings.pkl   # Saved face encodings
    └── models/
        └── person.py       # Data model for a person

1. Clone the repository:

   git clone https://github.com/muhit-khan/live_face_detection
   cd live_face_detection

2. Create and activate a virtual environment:

   python -m venv .venv
   source .venv/Scripts/activate  # On Windows
   # source .venv/bin/activate  # On macOS/Linux

3. Install the dependencies:

   pip install -r requirements.txt

   Note: Installing dlib can be tricky. You might need to install cmake first. If you face issues, please refer to the official installation guides for dlib and face_recognition.

To add a new person to the recognition system, you need to add one or more photos of them to the source_images directory.

• Image Naming Convention: The script extracts the person's name from the image filename. The name is the part of the filename before the first underscore (_).

  • Correct Naming:

    • Elon_Musk_1.jpg -> Name: Elon
    • Bill_Gates_2.png -> Name: Bill
    • Jeff_Bezos_3.jpeg -> Name: Jeff

  • Incorrect Naming:

    • Elon Musk 1.jpg (contains spaces)
    • Bill-Gates-2.png (uses hyphens instead of underscores)

  The text after the first underscore is ignored, so you can use it to number the images or add other information.

• Supported formats: The supported image formats are .jpg, .jpeg, .png, .bmp, and .tiff.

After adding new photos, you need to run the face encoding process. This will scan the source_images directory, detect faces in the new photos, and save their encodings to the src/face_encodings/encodings.pkl file.

Once the faces are encoded, you can start the live recognition process. This will open a window with your webcam feed and display the names of any recognized individuals.

There are two ways to run the functionalities of this project:

The easiest way to run the project is to use the run.sh script, which provides a simple command-line menu.

1. Open a terminal in the project's root directory.
2. Run the script:

   bash run.sh

3. Choose an option from the menu:
   • 1. Run Live Recognition: Starts the live face recognition.
   • 2. Encode Faces from Source Images: Encodes the faces from the source_images directory.
   • 3. Exit: Exits the script.

You can also run the Python scripts directly from the command line. Make sure your virtual environment is activated.

• To encode faces:

  python -m src.encode_faces

• To run live recognition:

  python -m src.live_recognition

• click==8.2.1
• cmake==4.1.0
• colorama==0.4.6
• dlib
• face-recognition==1.2.3
• face-recognition-models
• iniconfig==2.1.0
• numpy==2.2.6
• opencv-python==4.12.0.88
• packaging==25.0
• pillow==11.3.0
• pluggy==1.6.0
• Pygments==2.19.2
• pytest==8.4.2
• setuptools==80.9.0

Contributions are welcome! Please feel free to submit a pull request.

This project is licensed under the MIT License - see the LICENSE file for details.

Copyright (c) 2025 MUHIT KHAN