人脸检测最新文章汇总（基于深度学习方法）

 

导言
人脸检测是目前所有目标检测子方向中被研究的最充分的问题之一，它在安防监控，人证比对，人机交
互，社交和娱乐等方面有很强的应用价值，也是整个人脸识别算法的第一步。在本文中，SIGAI将和大
家一起回顾人脸检测算法的整个发展历史。

人脸检测综述-来自知乎

问题描述

人脸检测的目标是找出图像中所有的人脸对应的位置，算法的输出是人脸外接矩形在图像中的坐标，可能还包括姿态如倾斜角度等信息。下面是号称世界上最大的自拍照人脸检测结果（2019CVPR文章）：

 人脸检测特点：

虽然人脸的结构是确定的，由眉毛、眼睛、鼻子和嘴等部位组成，近似是一个刚体，但由于姿态和表情的变化，不同人的外观差异，光照，遮挡的影响，准确的检测处于各种条件下的人脸是一件相对困难的事情。

难点：

人脸检测算法要解决以下几个核心问题：

 人脸可能出现在图像中的任何一个位置
 人脸可能有不同的大小
 人脸在图像中可能有不同的视角和姿态
 人脸可能部分被遮挡

评价一个人脸检测算法好坏的指标是检测率和误报率。我们将检测率定义为： 

算法要在检测率和误报率之间做平衡，理想的情况是有高检测率，低误报率。

 基于深度学习的方法在FDDB上基本饱和了，是时候抛出一个新的benchmark了，即WIDERFace。

FDDB：

2845张图片中的5171张脸

标准人脸检测评测集

链接

WIDERFace：

32203张图片中的393703张脸

标准人脸检测评测集

地址

 基于WIDER Face的最新文章（包括2019CVPR文章）：

• Multiscale Cascade CNN - S. Yang, P. Luo, C. C. Loy, X. Tang. WIDER FACE: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016. 
  [Val PR Curves] [Test PR Curves]
• Two-stage CNN - S. Yang, P. Luo, C. C. Loy, X. Tang. WIDER FACE: IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016. 
  [Val PR Curves] [Test PR Curves]
• ACF-WIDER - B. Yang, J. Yan, Z. Lei and S. Z. Li. Aggregate channel features for multi-view face detection.. International Joint Conference on Biometrics, 2014. 
  [Val PR Curves] [Test PR Curves]
• Faceness-WIDER - S. Yang, P. Luo, C. C. Loy, X. Tang. From Facial Parts Responses to Face Detection: A Deep Learning Approach. IEEE International Conference on Computer Vision (ICCV), 2015. 
  [Val PR Curves] [Test PR Curves]
• Multitask Cascade CNN - K Zhang, Z Zhang, Z Li, Y Qiao. Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks. arXiv preprint arXiv:1604.02878, 2016. 
  [Val PR Curves] [Test PR Curves]
• CMS-RCNN - C Zhu, Y Zheng, K Luu, M Savvides. CMS-RCNN: Contextual Multi-Scale Region-based CNN for Unconstrained Face Detection. arXiv preprint arXiv:1606.05413, 2016. 
  [Val PR Curves] [Test PR Curves]
• LDCF+ - E. Ohn-Bar and M. Trivedi. To Boost or Not to Boost? On the Limits of Boosted Trees for Object Detection. International Conference on Pattern Recognition, (ICPR), 2016. 
  [Val PR Curves] [Test PR Curves]
• HR - P. Hu, D. Ramanan. Finding Tiny Faces. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017. 
  [Val PR Curves] [Test PR Curves]
• Face R-CNN - H. Wang, Z. Li, X. Ji, Y. Wang. Face R-CNN. arXiv preprint arXiv:1706.01061, 2017. 
  [Val PR Curves] [Test PR Curves]
• ScaleFace - S. Yang, Y. Xiong, C. C. Loy, X. Tang. Face Detection through Scale-Friendly Deep Convolutional Networks. arXiv preprint arXiv:1706.02863, 2017. 
  [Val PR Curves] [Test PR Curves]
• SSH - M. Najibi, P. Samangouei, R. Chellappa, L. Davis. SSH: Single Stage Headless Face Detector. IEEE International Conference on Computer Vision (ICCV), 2017. 
  [Val PR Curves] [Test PR Curves]
• SFD - S. Zhang, X. Zhu, Z. Lei, H. Shi, X. Wang, S. Z. Li. S³FD: Single Shot Scale-invariant Face Detector. IEEE International Conference on Computer Vision (ICCV), 2017. 
  [Val PR Curves SFD] [Test PR Curves SFD] 
  [Val PR Curves SFD-C] [Test PR Curves SFD-C]
• Face R-FCN - Y. Wang, X. Ji, Z. Zhou, H. Wang, Z. Li. Detecting Faces Using Region-based Fully Convolutional Networks. arXiv preprint arXiv:1709.05256, 2017. 
  [Val PR Curves] [Test PR Curves]
• MSCNN - Z. Cai, Q. Fan, R. Feris, N. Vasconcelos . A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection, European Conference on Computer Vision (ECCV), 2016. 
  [Val PR Curves] [Test PR Curves]
• FAN - J. Wang, Y. Yuan, G. Yu Face Attention Network: An Effective Face Detector for the Occluded Faces. arXiv preprint arXiv:1711.07246, 2017. 
  [Val PR Curves] [Test PR Curves]
• Zhu. et al. - C. Zhu, R. Tao, K. Luu, M. Savvides Seeing Small Faces from Robust Anchor's Perspective. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018. 
  [Val PR Curves] [Test PR Curves]
• PyramidBox - X. Tang, Daniel K. Du, Z. He, J. Liu PyramidBox: A Context-assisted Single Shot Face Detector. arXiv preprint arXiv:1803.07737, 2018. 
  [Val PR Curves] [Test PR Curves]
• FDNet - C. Zhang, X. Xu, D. Tu Face Detection Using Improved Faster RCNN. arXiv preprint arXiv:1802.02142, 2018. 
  [Val PR Curves] [Test PR Curves]
• SRN - C. Chi, S. Zhang, J. Xing, Z. Lei, S. Z. Li Selective Refinement Network for High Performance Face Detection. arXiv preprint arXiv:1809.02693, 2018. 
  [Val PR Curves] [Test PR Curves]
• FANet - J. Zhang, X. Wu, J. Zhu, Steven C.H. Hoi Feature Agglomeration Networks for Single Stage Face Detection. arXiv preprint arXiv:1712.00721, 2018. 
  [Val PR Curves] [Test PR Curves]
• DSFD - J. Li, Y. Wang, C. Wang, Y. Tai, J. Qian, J. Yang, C. Wang, J. Li, F. Huang DSFD: Dual Shot Face Detector. arXiv preprint arXiv:1810.10220, 2018. 
  [Val PR Curves] [Test PR Curves]
• DFS - W. Tian, Z. Wang, H. Shen, W. Deng, B. Chen, X. Zhang Learning Better Features for Face Detection with Feature Fusion and Segmentation Supervision. arXiv preprint arXiv:1811.08557, 2018. 
  [Val PR Curves] [Test PR Curves]
• VIM-FD - Y. Zhang, X. Xu, X. Liu Robust and High Performance Face Detector. arXiv preprint arXiv:1901.02350, 2019. 
  [Val PR Curves] [Test PR Curves]
• ISRN - S. Zhang, R. Zhu, X. Wang, H. Shi, T. Fu, S. Wang, T. Mei, Stan Z. Li Improved Selective Refinement Network for Face Detection. arXiv preprint arXiv:1901.06651, 2019. 
•