人脸检测研究2015最新进展(持续更新)

曾经终败给现在 2022-08-18 11:54 162阅读 0赞

搜集整理了2004~2015性能最好的人脸检测的部分资料，欢迎交流和补充相关资料。

转自：http://www.cvrobot.net/latest-progress-in-face-detection-2015/

1：人脸检测性能

### 1.1 人脸检测测评 ###

目前有两个比较大的人脸测评网站：

1：Face Detection Data Set and Benchmark（FDDB）

网址：[http://vis-www.cs.umass.edu/fddb/results.html][http_vis-www.cs.umass.edu_fddb_results.html]

FDDB是由马萨诸塞大学计算机系维护的一套公开数据库，为来自全世界的研究者提供一个标准的人脸检测评测平台，其中涵盖在自然环境下的各种姿态的人脸；该校还维护了LFW等知名人脸数据库供研究者做人脸识别的研究。作为全世界最具权威的人脸检测评测平台之一，FDDB使用Faces in the Wild数据库中的包含5171张人脸的2845张图片作为测试集，而其公布的评测集也代表了人脸检测的世界最高水平。

FDDB更新更及时一些，所以本文的资料还是主要参考的FDDB。

2：Fine-grained evaluation of face detection in the wild

网址：[http://www.cbsr.ia.ac.cn/faceEvaluation/results.html][http_www.cbsr.ia.ac.cn_faceEvaluation_results.html]

该测试网站是由李子青老师的研究组创立和维护的，其性能评估更细致，分析不同分辨率、角度、性别、年龄等条件下的算法准确率。该测试集更新没有FDDB及时。

### 1.2 Suevey ###

1）2010年微软zhang cha和张正友撰写的人脸检测的综述报告

[\[MSR-TR-2010\] A\_survey\_of\_recent\_advances\_in\_face\_detection][MSR-TR-2010_ A_survey_of_recent_advances_in_face_detection]

2）Stefanos Zafeiriou, Cha Zhang和张正友撰写了最新的人脸检测的综述paper，将出版在2016年的《Computer Vision and Image Understanding》

[\[CVIU 2015\] A Survey on Face Detection in the wild past, present and future][CVIU 2015_ A Survey on Face Detection in the wild past_ present and future]

最新性能总结如下：

[![A Survey on Face Detection in the wild past, present and future][A Survey on Face Detection in the wild past_ present and future]][A Survey on Face Detection in the wild past_ present and future_A Survey on Face Detection in the wild past_ present and future]

1）在过去的10年人脸检测的性能已经有了激动人心的提升。  
 2）这些引人注目的性能提升，主要还是得益于将Viala-Jones的boosting和鲁棒性的特征相组合。  
 3）始终有15~20%的性能Gap，即使允许一个相对较大的FP(大约1000）,始终有15~10%的人脸无法被检测到。需要特别指出的是这些Gap主要是由于是失焦的人脸（比如模糊的人脸）。  
 4）在这个Benchmark中，最好的基于boosting技术和最好的基于DPM的技术是比较接近的。当然最好的技术还是boosting和DPM组合在一起的性能。（这个就是指的[\[ECCV 2014\] Joint Cascade Face Detection and Alignment][ECCV 2014_ Joint Cascade Face Detection and Alignment]）

### 1.4 有关人脸检测指标 ###

如果对于人脸检测指标不是很熟悉，可以参考[http://www.cvrobot.net/recall-precision-false-positive-false-negative/][http_www.cvrobot.net_recall-precision-false-positive-false-negative]

[![recall-precision-false-positive-false-negative-1][]][recall-precision-false-positive-false-negative-1]

## 2. 2014的进展 ##

1：Joint Cascade Face Detection and Alignment. ECCV 2014. D. Chen, S. Ren, Y. Wei, X. Cao, J. Sun.  
 Paper：[\[ECCV 2014\] Joint Cascade Face Detection and Alignment][ECCV 2014_ Joint Cascade Face Detection and Alignment]

中文介绍：[联合人脸检测、校准算法介绍][Link 1]

2：The fastest deformable part model for object detection J. Yan, Z. Lei, L. Wen, S. Z. Li,

paper：[\[CVPR 2014\] The Fastest Deformable Part Model for Object Detection][CVPR 2014_ The Fastest Deformable Part Model for Object Detection]

3：Face detection without bells and whistles. ECCV 2014. M. Mathias, R. Benenson, M. Pedersoli and L. Van Gool.  
 Paper：[\[ECCV 2014\] Face detection without bells and whistles.][ECCV 2014_ Face detection without bells and whistles.]

project：[http://markusmathias.bitbucket.org/2014\_eccv\_face\_detection/][http_markusmathias.bitbucket.org_2014_eccv_face_detection]

Code：[https://bitbucket.org/rodrigob/doppia][https_bitbucket.org_rodrigob_doppia]

Talk： [http://videolectures.net/eccv2014\_mathias\_face\_detection/][http_videolectures.net_eccv2014_mathias_face_detection] （不错的报告）

Slide：[eccv2014\_mathias\_face\_detection\_01][eccv2014_mathias_face_detection_01]

[![Face detection without bells and whistles][]][Face detection without bells and whistles 1]

4：A Method for Object Detection Based on Pixel Intensity Comparisons Organized in Decision Trees. CoRR 2014. N. Markus, M. Frljak, I. S. Pandzic, J. Ahlberg and R. Forchheimer.  
 Code：[https://github.com/nenadmarkus/pico][https_github.com_nenadmarkus_pico]

Paper：[Object Detection with Pixel Intensity Comparisons Organized in Decision Trees][]

实时人脸检测视频Demo：

5：Aggregate channel features for multi-view face detection.. B. Yang, J. Yan, Z. Lei and S. Z. Li.

Paper：[\[IJCB 2014\] Aggregate channel features for multi-view face detection][IJCB 2014_ Aggregate channel features for multi-view face detection]

## 3. 2015的最新进展 ##

6：A Convolutional Neural Network Cascade for Face Detection. H. Li , Z. Lin , X. Shen, J. Brandt and G. Hua.   
 paper：[\[CVPR2015\] A Convolutional Neural Network Cascade for Face Detection][CVPR2015_ A Convolutional Neural Network Cascade for Face Detection]

7：Multi-view Face Detection Using Deep Convolutional Neural Networks. S. S. Farfade, Md. Saberian and Li-Jia Li

这是yahoo的人脸检测

Paper：[\[ICMR 2015\] Multi-view Face Detection Using Deep Convolutional Neural Networks][ICMR 2015_ Multi-view Face Detection Using Deep Convolutional Neural Networks]

News：[The Face Detection Algorithm Set to Revolutionize Image Search][]

[![\[ICMR 2015\] Multi-view Face Detection Using Deep Convolutional Neural Networks-1][ICMR 2015_ Multi-view Face Detection Using Deep Convolutional Neural Networks-1]][ICMR 2015_ Multi-view Face Detection Using Deep Convolutional Neural Networks-1]

8：Face Detection with a 3D Model. A. Barbu, N. Lay, G. Gramajo.  
 paper：[Face Detection with a 3D Model][]

结论：The 3D proposals are not perfectly aligned with the face keypoints, which results in a reduced accuracy in the high precision/very low false positive regime compared to other state of the art methods. However, in the regime of at least 0.1 false positives per image, it outperforms the cascade-based state of the art methods.

[http_vis-www.cs.umass.edu_fddb_results.html]: http://vis-www.cs.umass.edu/fddb/results.html
[http_www.cbsr.ia.ac.cn_faceEvaluation_results.html]: http://www.cbsr.ia.ac.cn/faceEvaluation/results.html
[MSR-TR-2010_ A_survey_of_recent_advances_in_face_detection]: http://www.cvrobot.net/wp-content/uploads/2015/06/MSR-TR-2010-A_survey_of_recent_advances_in_face_detection.pdf
[CVIU 2015_ A Survey on Face Detection in the wild past_ present and future]: http://www.cvrobot.net/wp-content/uploads/2015/06/CVIU-2015-A-Survey-on-Face-Detection-in-the-wild-past-present-and-future.pdf
[A Survey on Face Detection in the wild past_ present and future]: http://www.cvrobot.net/wp-content/uploads/2015/06/A-Survey-on-Face-Detection-in-the-wild-past-present-and-future-1024x359.png
[A Survey on Face Detection in the wild past_ present and future_A Survey on Face Detection in the wild past_ present and future]: /images/20220731/be82d19a9ecf4e64bb2800a477f544f3.png
[ECCV 2014_ Joint Cascade Face Detection and Alignment]: http://www.cvrobot.net/wp-content/uploads/2015/06/ECCV-2014-Joint-Cascade-Face-Detection-and-Alignment1.pdf
[http_www.cvrobot.net_recall-precision-false-positive-false-negative]: http://www.cvrobot.net/recall-precision-false-positive-false-negative/
[recall-precision-false-positive-false-negative-1]: /images/20220731/8360d9de853e4392961d7089a372eee2.png
[Link 1]: http://www.cvrobot.net/joint-cascade-face-detection-and-alignment/
[CVPR 2014_ The Fastest Deformable Part Model for Object Detection]: http://www.cvrobot.net/wp-content/uploads/2015/06/CVPR-2014-The-Fastest-Deformable-Part-Model-for-Object-Detection.pdf
[ECCV 2014_ Face detection without bells and whistles.]: http://www.cvrobot.net/wp-content/uploads/2015/06/ECCV-2014-Face-detection-without-bells-and-whistles..pdf
[http_markusmathias.bitbucket.org_2014_eccv_face_detection]: http://markusmathias.bitbucket.org/2014_eccv_face_detection/
[https_bitbucket.org_rodrigob_doppia]: https://bitbucket.org/rodrigob/doppia
[http_videolectures.net_eccv2014_mathias_face_detection]: http://videolectures.net/eccv2014_mathias_face_detection/
[eccv2014_mathias_face_detection_01]: http://www.cvrobot.net/wp-content/uploads/2015/06/eccv2014_mathias_face_detection_01.pdf
[Face detection without bells and whistles]: /images/20220731/f94f867a43ed4257a7104baacf857717.png
[Face detection without bells and whistles 1]: http://www.cvrobot.net/wp-content/uploads/2015/06/Face-detection-without-bells-and-whistles.png
[https_github.com_nenadmarkus_pico]: https://github.com/nenadmarkus/pico
[Object Detection with Pixel Intensity Comparisons Organized in Decision Trees]: http://www.cvrobot.net/wp-content/uploads/2015/06/Object-Detection-with-Pixel-Intensity-Comparisons-Organized-in-Decision-Trees.pdf
[IJCB 2014_ Aggregate channel features for multi-view face detection]: http://www.cvrobot.net/wp-content/uploads/2015/06/IJCB-2014-Aggregate-channel-features-for-multi-view-face-detection.pdf
[CVPR2015_ A Convolutional Neural Network Cascade for Face Detection]: http://www.cvrobot.net/wp-content/uploads/2015/06/CVPR2015-A-Convolutional-Neural-Network-Cascade-for-Face-Detection.pdf
[ICMR 2015_ Multi-view Face Detection Using Deep Convolutional Neural Networks]: http://www.cvrobot.net/wp-content/uploads/2015/06/ICMR-2015-Multi-view-Face-Detection-Using-Deep-Convolutional-Neural-Networks.pdf
[The Face Detection Algorithm Set to Revolutionize Image Search]: http://www.technologyreview.com/view/535201/the-face-detection-algorithm-set-to-revolutionize-image-search/
[ICMR 2015_ Multi-view Face Detection Using Deep Convolutional Neural Networks-1]: /images/20220731/66cafa77b58541a6b230e10093c9af45.png
[Face Detection with a 3D Model]: http://www.cvrobot.net/wp-content/uploads/2015/06/Face-Detection-with-a-3D-Model.pdf