dlib Face Recognition

About

dlib is a modern C++ toolkit containing machine learning algorithms and tools for creating complex software. Its face recognition module offers a simple yet powerful pipeline including HOG-based face detection, CNN face detection, 68-point facial landmark prediction, and a face recognition model with 99.38% accuracy on LFW benchmark.

Key Features

• HOG face detector
• CNN face detector
• 68-point landmark predictor
• Face recognition (99.38% LFW)
• Object tracking
• Python bindings

Use Cases

• Academic research
• Face attendance systems
• Photo tagging
• Security applications
• Face clustering