计算机科学 ›› 2014, Vol. 41 ›› Issue (7): 301-305.doi: 10.11896/j.issn.1002-137X.2014.07.062

• 图形图像与模式识别 • 上一篇    下一篇

基于SURF与Hough森林的人脸检测研究

严明君,项俊,罗艳,侯建华   

  1. 中南民族大学智能无线通信湖北省重点实验室 武汉430074;中南民族大学电子信息工程学院 武汉430074;华中科技大学自动化学院 武汉430074;中南民族大学智能无线通信湖北省重点实验室 武汉430074;中南民族大学电子信息工程学院 武汉430074;中南民族大学智能无线通信湖北省重点实验室 武汉430074;中南民族大学电子信息工程学院 武汉430074
  • 出版日期:2018-11-14 发布日期:2018-11-14
  • 基金资助:
    本文受国家自然科学基金项目(61141010),武汉市科技供需对接计划项目(201051824575),湖北省自然科学基金项目(2012FFA113),中南民族大学中央高校基本科研业务费专项资金项目(CZY13033)资助

Face Detection Based on SURF and Hough Forests

YAN Ming-jun,XIANG Jun,LUO Yan and HOU Jian-hua   

  • Online:2018-11-14 Published:2018-11-14

摘要: 为实现复杂场景中的人脸检测与定位,提出了一种基于快速鲁棒特征SURF与Hough森林的人脸检测算法。采用SURF局部特征构建Hough森林分类器,每个叶子节点存储类别信息与特征点到达目标中心的偏移量,在图像局部外观与Hough投票之间建立映射关系,生成有监督的判别式的码本,获得可靠的概率Hough投票,以此预测目标中心位置,提高了检测精度。与此同时,采用SURF局部特征提取图像兴趣点有助于减小计算量、加快检测速度。实验证明了所提算法的有效性。

关键词: SURF,Hough森林,决策树,训练分类器,概率Hough投票 中图法分类号TP391文献标识码A

Abstract: In order to realize the face detection and localization in complicated scenes,this paper presented an algorithm for face detection based on SURF (speeded up robust features) and Hough forests.SURF local features were adopted to construct a Hough forest classifier.Each leaf node stored the class information as well as the offsets from the locations of interest points to the centroid of the object,and the mapping relationship between local appearances of images and their Hough votings were established.The supervised and discriminative codebook was generated,which was used to estimate the object’s location via a probabilistic Hough voting,thereby improving the detection precision.Meanwhile,the algorithm reduced the computation and made the detection faster by using SURF local features.Experimental results demonstrated the efficiency of the proposed algorithm.

Key words: SURF,Hough forest,Decision tree,Training classifier,Probabilistic Hough voting

[1] Yang D W,Chen C W,Chang C H,et al.Face detection architecture design using hybrid skin color detection and cascade of classifiers[C]∥2012IEEE Asia Pacific Conference on Circuits and Systems (APCCAS). Taipei:IEEE,2012:543-546
[2] Yang M H,Kriegman D,Ahuja N.Detecting faces in images:A survey[J].IEEE Trans.Pattern Analysis and Machine Intelligence,2002,24:34-58
[3] Zhu X,Ramanan D.Face detection,pose estimation,and landmark localization in the wild[C]∥Computer Vision and Pattern Recognition,2012IEEE Conference on Providence.RI USA:IEEE,2012:2879-2886
[4] Becker C,Rigamonti R,Lepetit V,et al.KernelBoost:SupervisedLearning of Image Features For Classification[R].Lausanne,Swiss:EPFL,2013
[5] Lowe D.Distinctive image features from scale-invariant keypoint[J].International Journal of Computer Vision,2004,60(2):91-110
[6] Mikolajczyk K,Schmid C.A performance evaluation of local de-scriptors[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(10):1615-1630
[7] Ke Y,Sukthankar R.PCA-SIFT:a more distinctive representation for local image descriptors[C]∥IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Wa-shington D C,USA:IEEE,2004:506-513
[8] Bay H,Ess A,Tuytelaars T.SURF:Speeded up robust features[J].Computer Vision and Image Understanding,2008,100(3):346-359
[9] Fergus R,Perona P,Zisserman A.Weakly supervised scale-invariant learning of models for visual recognition[J].Interna-tional Journal of Computer Vision,2008,71(3):273-303
[10] Leibe B,Leonardis A,Schiele B.Robust object detection with interleaved categorization and segmentation[J].International Journal of Computer Vision,2008,77(1-3):259-289
[11] Bosch A,Zisserman A,Muoz X.Image classification using random forests and ferns[C]∥IEEE 11th International Conference on Computer Vision,2007(ICCV 2007).Rio de Janeiro,Brazil,IEEE,2007:1-8
[12] Borenstein E,Ullman S.Class-specific,top-down segmen-tation[C]∥the 7th European Conference on Computer Vision.Copenhagen,Denmark,ECCV,2002:109-122
[13] Hinterstoisser S,Cagniart C,Ilic S,et al.Gradient responsemaps for real-time detection of textureless objects[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(5):876-888
[14] Breiman L.Random Forest[J].Machine Learning,2001,45(1):5-32
[15] Rodriguez-Galiano V F,Ghimire B,Rogan J,et al.An assess-ment of the effectiveness of a random forest classifier for land-cover classification[J].ISPRS Journal of Photogrammetry and Remote Sensing,2012,67:93-104
[16] Gall J,Lempitsky V.Class-specific hough forests for object detection[C]∥IEEE Conference on Computer Vision and Pattern Recognition.Miami,FL,USA,CVPR,2009:1022-1029
[17] Dalal N,Triggs B.Histograms of oriented gradients for human detection[C]∥IEEE Computer Society Conference on Computer Vision and Pattern Recognition.San Diego,CVPR,2005,1:886-893
[18] Opelt A,Pinz A,Zisserman A.Learning an alphabet of shape and appearance for multi-class object detection [J].IJCV,2008,80(1):16-44
[19] 陈艺虾,孙权森,徐焕宇,等.SURF算法和RANSAC算法相结合的遥感图像匹配方法[J].计算机科学,2012,6(9):822-828
[20] Viola P,Jones M.Rapid object detection using a boosted cascade of simple features[C]∥IEEE Conference on Computer Vision and Pattern Recognition.Kauai Hawaii USA,IEEE Computer Society,2001:511-518
[21] Mikolajczyk K,Schmid C.Indexing based on scale invariant interest points[C]∥IEEE 8th International Conference on Computer Vision,2001(ICCV 2001).Vancouver,BC,Canada:IEEE,2001,1:525-531
[22] Simard Y,Bottou L,Haffner P,et al.Boxlets:a fast convolution algorithm for signal processing and neural networks[J].Advances in Neural Information Processing Systems,1999:571-577
[23] Brown M,Lowe D.Invariant features from interest point groups[C]∥British Machine Vision Conference.Cardiff,Wales,BMVC,2002:20-28
[24] Liu H,Cui W,Ding R.Robust Hand Tracking with Hough Forest and Multicue Flocks of Features[C]∥the 8th International Symposium on Visual Computing 2012(ISVC).Rethymnon,Crete,Greece,2012:458-467

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!