FaceAuth

FaceAuth is a user authentication system based on face recognition, featuring a modern graphical interface and seamless integration with Linux systems (e.g., PAM). The application allows you to log in or authorize privileged actions (such as sudo) using your face instead of a typed password, or as an additional security layer.

Remember:
FaceAuth uses 2D face recognition based on a standard webcam image. This approach is convenient and user-friendly, but it is not foolproof or as secure as advanced biometric systems (such as 3D or infrared recognition). The system may be vulnerable to sophisticated spoofing attacks (e.g., high-quality photos or videos). For critical security scenarios, always use FaceAuth as a supplement to strong passwords, not as the sole protection.

Features

• Add and remove your own face templates (stored locally)
• Real-time face recognition for the user
• System login integration (PAM) via a custom daemon and C module
• User-friendly graphical interface (PyQt5)
• Support for multiple face templates per user

How it works

1. Adding a face – The user positions their face in front of the camera and saves a template, which is processed by dlib models and stored as a feature vector.
2. Recognition – During login, the application compares the current camera image with saved templates.
3. PAM integration – The C module communicates with the daemon, which performs face recognition and returns the result to the system.

Technologies used

• Python 3 – main application language
• PyQt5 – graphical user interface
• OpenCV – camera handling and image processing
• dlib – face detection, landmark extraction, and feature vector generation
• NumPy – numerical operations and feature vector handling
• PAM (Pluggable Authentication Modules) – Linux login integration (optional)
• C – intermediary module for PAM

Models used

• 68_face_landmarks_model_v2.dat – custom-trained dlib model for detecting 68 facial landmarks, required for proper feature extraction.
• dlib_face_recognition_resnet_model_v1.dat – dlib model based on ResNet, generating a unique face feature vector for comparison and recognition.

Requirements

• Python 3.7+
• dlib
• OpenCV
• PyQt5
• NumPy
• PAM (optional, for system integration)

Installation

1. Install the required libraries:

   pip install -r requirements.txt

   If you don't have a requirements.txt file, you can install the main dependencies manually:

   pip install dlib opencv-python PyQt5 numpy

2. Make sure the dlib models are present in the project directory:

   • 68_face_landmarks_model_v2.dat
   • dlib_face_recognition_resnet_model_v1.dat

3. Create your face template(s):

   • Run the main application to add your face:

     python3 main.py

   • Use the GUI to add and save your face template(s). These will be stored locally for your user.

4. Set up the PAM daemon:

   • Add pam-module-deamon.py to autostart (e.g., as a systemd service or a background process at login), so it runs in the background and listens for authentication requests.

5. Configure PAM:

   • Edit the PAM configuration file for the service you want to protect, for example for sudo:

     sudo nano /etc/pam.d/sudo

   • Add the following line at the top:

     auth sufficient pam_exec.so stdout /path/to/face-auth/face-auth
     

     Replace /path/to/face-auth with the path to your compiled C module that communicates with the daemon.

6. (Optional) Compile the C module:

   • If you use the provided C intermediary for PAM:

     gcc -o face-auth face_auth.c

License

Project released under the MIT license.