Results

We present two types of scoring the detections in an image: discrete score, and continuous score. Under the former criterion, if the ratio of the intersection of a detected region with an annotated face region is greater than 0.5, a score of 1 is assigned to the detected region, and 0 otherwise. On the other hand, for the second criterion, this ratio is used as the score for the detected region. Further details for the evaluation procedure can be found in the FDDB technical report.

Generating performance curves

Evaluation Code:


Code for matching detections and annotations and computing the resulting scores to generate the performance curves.

Source code: [tar-gzipped]
[last updated: May 06, 2010 1440 EDT]

The code has been successfully compiled and tested on Mac OS X, Linux, and Windows. Please check end of FAQ page for problems in running the evaluation code.


Comparison Code:


Gnuplot scripts for generating the ROC curves are provided below.
[last updated: April 1, 2020 1347 EST]


The curve plotting scripts and the ROC curves in a single tar file. [tar-gzipped]
[last updated: May 17, 2015 1304 EDT]

The latest results may not be updated always in the downloaded tar file. In that case, please copy the latest ROC curves individually into your local compareROC/rocCurves/ folder and re-run the plotting scripts.


Reporting results

 Please send the text files for the ROC curves to Aruni RoyChowdhury. See the Methods section below for sample output files.


Results for published methods

 Here, we present the results for the following face detection systems (Oldest first):
[last updated: April 1, 2020 1335 EST]
  1. K. Mikolajczyk, C. Schmidt and A. Zisserman. Human detection based on a probabilistic assembly of robust part detectors. ECCV 2004.
    [ DiscROC, ContROC ]
  2. W. Kienzle, G. Bakir, M. Franz and B. Scholkopf Face Detection - Efficient and Rank Deficient. Advances in Neural Information Processing Systems, 2005.
    [ DiscROC, ContROC ]
  3. OpenCV implementation of Viola-Jones face detector.
    [ DiscROC, ContROC ] -- updated on June 7, 2010.
  4. B. Subburaman Venkatesh and S. Marcel. Fast Bounding Box Estimation based Face Detection. ECCV Workshop on Face Detection, 2010.
    [ DiscROC, ContROC ]
  5. V. Jain and E. Learned-Miller. Online Domain Adaptation of a Pre-Trained Cascade of Classifiers. CVPR 2011.
    [ DiscROC, ContROC ]
  6. J. Li, T. Wang and Y. Zhang. Face Detection using SURF Cascade. ICCV 2011 BeFIT workshop.
    [ DiscROC, ContROC ]
  7. M. Koestinger, P. Wohlhart, P. M. Roth and H. Bischof. Robust Face Detection by Simple Means. DAGM 2012 CVAW workshop.
    [ DiscROC, ContROC ]
  8. S. Segui, M. Drozdzal, P. Radeva and J. Vitri. An Integrated Approach to Contextual Face Detection. ICPRAM 2012.
    [ DiscROC, ContROC ]
  9. X. Zhu and D. Ramanan. Face detection, pose estimation and landmark localization in the wild. CVPR 2012. a
    [ DiscROC, ContROC ]
  10. X. Shen, Z. Lin, J. Brandt and Y. Wu. Detecting and Aligning Faces by Image Retrieval. CVPR 2013. (legend: XZJY)
    [ DiscROC, ContROC ]
  11. H. Li, G. Hua, Z. Lin, J. Brandt and J. Yang. Probabilistic Elastic Part Model for Unsupervised Face Detector Adaptation. ICCV 2013. (legend: PEP-Adapt)
    [ DiscROC, ContROC ]
  12. J. Li and Y. Zhang. Learning SURF cascade for fast and accurate object detection. CVPR 2013. (legend: SURF frontal/multiview)

  13. H. Li, Z. Lin, J. Brandt, X. Shen and G. Hua. Efficient Boosted Exemplar-based Face Detection. CVPR 2014. (legend: Boosted Exemplar)
    [ DiscROC, ContROC ]
  14. J. Yan, Z. Lei, L. Wen and S. Z. Li. The Fastest Deformable Part Model for Object Detection. CVPR 2014.
    [ DiscROC, ContROC ]
  15. D. Chen, S. Ren, Y. Wei, X. Cao, J. Sun. Joint Cascade Face Detection and Alignment. ECCV 2014. (legend: Joint Cascade)
    [ DiscROC, ContROC ]
  16. M. Mathias, R. Benenson, M. Pedersoli and L. Van Gool. Face detection without bells and whistles. ECCV 2014. (legend: HeadHunter)
    [ DiscROC, ContROC ]
  17. N. Markus, M. Frljak, I. S. Pandzic, J. Ahlberg and R. Forchheimer. A Method for Object Detection Based on Pixel Intensity Comparisons Organized in Decision Trees. CoRR 2014. (legend: Pico)
    [ DiscROC, ContROC ]
  18. B. Yang, J. Yan, Z. Lei and S. Z. Li. Aggregate channel features for multi-view face detection.. International Joint Conference on Biometrics, 2014. (legend: ACF-multiscale)
    [ DiscROC, ContROC ]
  19. H. Li , Z. Lin , X. Shen, J. Brandt and G. Hua. A Convolutional Neural Network Cascade for Face Detection. Computer Vision and Pattern Recognition (CVPR), 2015. (legend: CascadeCNN)
    [ DiscROC, ContROC ]
  20. S. S. Farfade, Md. Saberian and Li-Jia Li. Multi-view Face Detection Using Deep Convolutional Neural Networks. International Conference on Multimedia Retrieval (ICMR), 2015. (legend: DDFD)
    [ DiscROC, ContROC ]
  21. A. Barbu, N. Lay, G. Gramajo. Face Detection with a 3D Model.

  22. G. Ghiasi, C. Fowlkes. Occlusion Coherence: Detecting and Localizing Occluded Faces. Technical Report, June 2015 [arXiv:1506.08347]. (legend: MultiresHPM)
    (extension of CVPR2014 paper)
    [ DiscROC, ContROC ]
  23. S. Liao, A. Jain, S. Li. A Fast and Accurate Unconstrained Face Detector. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015. (legend: NPDFace)
    [ DiscROC, ContROC ]
  24. 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. (legend: Faceness)
    [ DiscROC, ContROC ]
  25. B. Yang, J. Yan, Z. Lei, S. Z. Li. Convolutional Channel Features. IEEE International Conference on Computer Vision (ICCV), 2015. (legend: CCF)
    [ DiscROC, ContROC ]
  26. R. Ranjan, V. M. Patel, R. Chellappa. A Deep Pyramid Deformable Part Model for Face Detection. IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS), 2015. (legend: DP2MFD)
    [ DiscROC, ContROC ]
  27. Vijay Kumar, Anoop Namboodiri, C V Jawahar. Visual Phrases for Exemplar Face Detection. IEEE International Conference on Computer Vision (ICCV), 2015.
    [ DiscROC, ContROC ]
  28. Kaipeng Zhang, Zhanpeng Zhang, Zhifeng Li. Joint Face Detection and Alignment using Multi-task Cascaded Convolutional Networks. IEEE Signal Processing Letters, vol. 23, no. 10, pp. 1499-1503, 2016.
    [ DiscROC, ContROC ]
  29. Rajeev Ranjan, Vishal M. Patel, Rama Chellappa. HyperFace: A Deep Multi-task Learning Framework for Face Detection, Landmark Localization, Pose Estimation, and Gender Recognition. ArXiv report, 2016. (legend: HyperFace)
    [ DiscROC, ContROC ]
  30. Huaizu Jiang and Erik Learned-Miller. Face Detection with the Faster R-CNN. ArXiv report, 2016. (legend: Faster RCNN)
    [ DiscROC, ContROC ]
  31. Yunzhu Li, Benyuan Sun, Tianfu Wu, Yizhou Wang. Face Detection with End-to-End Integration of a ConvNet and a 3D Model. European Conference on Computer Vision (ECCV), 2016. (legend: Conv3D)
    [ DiscROC, ContROC ]
  32. Jiahui Yu, Yuning Jiang, Zhangyang Wang, Zhimin Cao, Thomas Huang. UnitBox: An Advanced Object Detection Network Arxiv report, 2016. (legend: UnitBox)
    [ DiscROC, ContROC ]
  33. Shaohua Wan, Zhijun Chen, Tao Zhang, Bo Zhang, Kong-kat Wong Bootstrapping Face Detection with Hard Negative Examples Arxiv report, 2016. (legend: Xiaomi Inc.)
    [ DiscROC, ContROC ]
  34. MXNet, open source code and models, 2016. (legend: mxnet)
    [ DiscROC, ContROC ]
  35. Danai Triantafyllidou and Anastasios Tefas A Fast Deep Convolutional Neural Network for Face Detection in Big Visual Data . Advances in Big Data, 2016. (legend: FastCNN)
    [ DiscROC, ContROC ]
  36. 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, 2016. (legend: LDCF+)
    [ DiscROC, ContROC ]
  37. DeepIR: Face Detection using Deep Learning: An Improved Faster RCNN Approach, ArXiv 2017. (legend: DeepIR)
    [ DiscROC, ContROC ]
  38. Peiyun Hu, Deva Ramanan. Finding Tiny Faces. Computer Vision and Pattern Recognition (CVPR), 2017. (legend: HR-ER, HR)
  39. H. Wang, Z. Li, X. Ji, Y. Wang. Face R-CNN , ArXiv 2017. (legend: Face_R-CNN)
    [ DiscROC, ContROC ]
  40. S. Yang, Y. Xiong, C. C. Loy, X. Tang. Face Detection through Scale-Friendly Deep Convolutional Networks , ArXiv 2017. (legend: Scale-Face)
    [ DiscROC, ContROC ]
  41. S. Zhang, X. Zhu, Z. Lei, H. Shi, X. Wang and S. Z. Li. FaceBoxes: A CPU Real-time Face Detector with High Accuracy, International Joint Conference on Biometrics (IJCB), 2017. (legend: FaceBoxes)
    [ DiscROC, ContROC ]
  42. S. Zhang, X. Zhu, Z. Lei, H. Shi, X. Wang and S. Z. Li. S³FD: Single Shot Scale-invariant Face Detector, International Conference on Computer Vision (ICCV), 2017. (legend: SFD)
    [ DiscROC, ContROC ]
  43. Y. Liu, H. Li, J. Yan, F. Wei, X. Wang, X. Tang. Recurrent Scale Approximation for Object Detection in CNN, International Conference on Computer Vision (ICCV), 2017. (legend: RSA)
    [ DiscROC, ContROC ]
  44. Y. Wang, X. Ji, Z. Zhou, H. Wang, Z. Li. Detecting Faces Using Region-based Fully Convolutional Networks, ArXiv 2017. (legend: Face R-FCN)
    [ DiscROC, ContROC ]
  45. K. Zhang, Z. Zhang, H. Wang, Z. Li, Y. Qiao and W. Liu. Detecting Faces Using Inside Cascaded Contextual CNN, International Conference on Computer Vision (ICCV), 2017. (legend: ICC-CNN)
    [ DiscROC, ContROC ]
  46. X. Tang, D. K. Du, Z. He and J. Liu. PyramidBox: A Context-assisted Single Shot Face Detector, ArXiv report, 2018. (legend: PyramidBox)
    [ DiscROC, ContROC ]
  47. D. Triantafyllidou, P. Nousi and A. Tefas. Fast Deep Convolutional Face Detection in the Wild Exploiting Hard Sample Mining, Big Data Research, 2017. (legend: FD-CNN)
    [ DiscROC, ContROC ]
  48. J. Zhang, X. Wu, J. Zhu and S. C.H. Hoi. Feature Agglomeration Networks for Single Stage Face Detection, ArXiv report, 2018. (legend: FANet)
    [ DiscROC, ContROC ]
  49. J. Li, Y. Wang, C. Wang, Y. Tai, J. Qian, J. Yang, C. Wang, J. Li, F. Huang. DSFD: Dual Shot Face Detector, ArXiv report, 2018. (legend: DSFD)
    [ DiscROC, ContROC ]
  50. L. Liu, G. Li, Y. Xie, Y. Yu, Q, Wang, L. Lin. Facial Landmark Machines: A Backbone-Branches Architecture with Progressive Representation Learning, IEEE Transactions on Multimedia, 2019. (legend: BBFCN)
    [ DiscROC, ContROC ]
  51. B. Zhang, J. Li, Y. Wang, Y. Tai, C. Wang, J. Li, F. Huang, Y. Xia, W. Pei, R. Ji. ASFD: Automatic and Scalable Face Detector, ArXiv, 2020. (legend: ASFD)
    [ DiscROC, ContROC ]

a. System trained on external data. The face_p146_small.mat model provided by the authors was used to evaluate on FDDB.

Performance curves (click to toggle zoom):
(a) Discrete Score

(b) Continuous Score



Results for unpublished methods

Here, we present the results for the following face detection systems:
[last updated: March 30, 2016 1030 EDT]

  1. IlluxTech frontal face detector.
    [ DiscROC, ContROC ]
  2. Olaworks face detector.
    [ DiscROC, ContROC ]
  3. Shenzhen University frontal face detector.
    [ DiscROC, ContROC ]
  4. TVPlay.cn face detector -- Shenzhen TVPlay technology Co., Ltd..
    [ DiscROC, ContROC ]
  5. Hisign face detector [link].
  6. Face++ face detectora [link].
    [ DiscROC, ContROC ]
  7. Shenzhen University face detector (2014) by Shiqi Yu.
    [ DiscROC, ContROC ]
  8. Tencent Best-Image [link]
    [ DiscROC, ContROC ]
  9. StradVision [link].
    [ DiscROC, ContROC ]
  10. Eyedea Recognition [link].
    [ DiscROC, ContROC ]
  11. Uni-Ubi [link].
    [ DiscROC, ContROC ]
  12. Intelligent Media Computing (IMC) Laboratory, Sun Yat-Sen University [link].
    [ DiscROC, ContROC ]
  13. Uni-Ubi (2015) [link].
    [ DiscROC, ContROC ]
  14. Linkface face detector [link].
    [ DiscROC, ContROC ]
  15. Baidu-IDL [link].
  16. PCI [link].
    [ DiscROC, ContROC ]
  17. Yunnan University.
    [ DiscROC, ContROC ]
  18. 360-NUS.
    [ DiscROC, ContROC ]
  19. CW-DNA, CloudWalk Inc [link].
  20. Beijing Faceall Technology Co., Ltd. [link].
  21. Color Reco [link].
    [ DiscROC, ContROC ]
  22. MS-FRCNN (K. Luu et al.)
    [ DiscROC, ContROC ]
  23. authenmetric.com
  24. MT-Face [link]
  25. Daream [link]
    [ DiscROC, ContROC ]
  26. Emotibot [link]
    [ DiscROC, ContROC ]
  27. Uniview [link]
    [ DiscROC, ContROC ]
  28. THU CV-AI Lab (Tsinghua University, Beijing)
    [ DiscROC, ContROC ]
  29. Tencent AI.v1
    [ DiscROC, ContROC ]
  30. Wisesoft b [link]
    [ DiscROC, ContROC ]
  31. ReadSense Ltd. [link]
  32. Glasssix [link]
    • Detection speed (CPU:i7-6700k RAM:8GB GPU:GTX titan x): 13FPS
    • Training data: widerface + celebA + private dataset

    [ DiscROC, ContROC ]
  33. NanYun Tech
    [ DiscROC, ContROC ]
  34. Ping An Technology AI lab
    [ DiscROC, ContROC ]
  35. Fuzhou Haijing Science and Technology Development Co.LTD [link]
    [ DiscROC, ContROC ]
  36. China Electronics Technology Cyber Security Co.,Ltd (CETCSC) [link]
    [ DiscROC, ContROC ]
  37. Wisers AI Lab [link]
    [ DiscROC, ContROC ]
  38. YY.inc (SSD-based)
    [ DiscROC, ContROC ]
  39. FABU Technology [link]
    [ DiscROC, ContROC ]

a. Not the same version as the detector used by the free Face++ API

b. Method Description:
  • Detection speed (CPU:i7-3770 RAM:8GB GPU:GTX1080): 16FPS
  • Training data: WIDER + AFLW


Performance curves (click to toggle zoom):
(a) Discrete Score
(b) Continuous Score