paper
github提供预训练model
先来看一下该文强大的背景：
2019.08.10: We achieved 2nd place at WIDER Face Detection Challenge 2019.

2019.04.30: Our Face detector (RetinaFace) obtains state-of-the-art results on the WiderFace dataset.

该文的主要贡献：

• Based on a single-stage design, we propose a novel pixel-wise face localisation method named Reti- naFace, which employs a multi-task learning strategy to simultaneously predict face score, face box, five fa- cial landmarks, and 3D position and correspondence of each facial pixel.
  • On the WIDER FACE hard subset, RetinaFace outper- forms the AP of the state of the art two-stage method (ISRN [67]) by 1.1% (AP equal to 91.4%).
  • On the IJB-C dataset, RetinaFace helps to improve Ar- cFace’s [11] verification accuracy (with TAR equal to 89.59% when FAR=1e-6). This indicates that better face localisation can significantly improve face recog- nition.
  • By employing light-weight backbone networks, Reti- naFace can run real-time on a single CPU core for a VGA-resolution image.
  • Extra annotations and code have been released to fa- cilitate future research.