Retina

What is Retina?

Retina is a Face Recognition application. It uses OpenCV and has "optimizied" settings for running on several of my machines.

Project Hierarchy

• data - Permanent and temporary data used for detection and recognition.
  • classifiers - Haar and LBP face detection classifiers.
  • faces - Face image sets for specific individuals.
    • LABEL - Sets for a specific individual.
      • raw - Raw, non-preprocessed, RGB images from an external source.
      • training - Preprocessed images ready for recognizer training.
  • recognizers - Face recognizer models.
• modules - Internal data models and helper functions.
• settings - Machine specific settings for video capture, detection, and recognition.
• tools - Various tools for working with Retina.

Dependencies

• Python 3.x
• OpenCV >= 3.2.0
• Numpy
• Pillow
• Tkinter (optional)

Recommended OpenCV dependencies

• Eigen3
• FFmpeg
• V4l / V4l2
• TBB

Installing OpenCV

To install OpenCV 3+ for Python 3.x, I've been following the install instructions located here. If you are working on a Fedora-based machine like I am, simply installing the RPM equivalents of the Step 1 packages and following the rest of the guide should be fine. You can find OpenCV and OpenCV_Contrib here and here.