ageitgey/face_recognition

Built on dlib's deep learning models achieving 99.38% accuracy on standard benchmarks, it provides face detection, landmark identification (eyes, nose, mouth, chin), and identity comparison through simple encoding and matching functions. Supports real-time processing via webcam integration and batch processing through both a Python API and dedicated command-line tools (`face_recognition` and `face_detection`) for folder-based recognition workflows.

56,193 stars and 234,099 monthly downloads. No commits in the last 6 months. Available on PyPI.