I selected 5 ready-made algorithms for face detection and compared them with each other by such metrics as Precision, Recall, IOU and time on the dataset I marked up. Results:
- https://pyimagesearch.com/2016/11/07/intersection-over-union-iou-for-object-detection/
- https://github.com/Linzaer/Ultra-Light-Fast-Generic-Face-Detector-1MB
- https://github.com/ageitgey/face_recognition
- https://habr.com/ru/company/ods/blog/328372/
- https://github.com/timesler/facenet-pytorch
- https://github.com/serengil/retinaface
- https://github.com/dlunion/DBFace