联合梯度直方图和局部二值模式特征的人体检测

黄炎; 范赐恩; 朱秋平; 张虎; 邓德祥

doi:null

您当前的位置：

首页 >

文章列表页 >

联合梯度直方图和局部二值模式特征的人体检测

信息科学 | 更新时间：2020-08-12

- 联合梯度直方图和局部二值模式特征的人体检测
- HOG-LBP pedestrian detection
- 光学精密工程 2013年21卷第4期页码：1047-1053
- 作者机构：
  
  武汉大学电子信息学院,湖北武汉,中国,430072
- 作者简介：
- 基金信息：
  
  面向非规则计算算法的FPGA逻辑映射优化()
- DOI：
  中图分类号： TP391
- 收稿日期：2012-09-25，
  
  修回日期：2012-11-06，
  
  网络出版日期：2013-04-20，
  
  纸质出版日期：2013-04-15
- 稿件说明：
移动端阅览
黄炎范赐恩朱秋平张虎邓德祥. 联合梯度直方图和局部二值模式特征的人体检测[J]. 光学精密工程, 2013,21(4): 1047-1053

HUANG Yan FAN Ci-en ZHU Qiu-ping ZHANG Hu DENG De-xiang. HOG-LBP pedestrian detection[J]. Editorial Office of Optics and Precision Engineering, 2013,21(4): 1047-1053
黄炎范赐恩朱秋平张虎邓德祥. 联合梯度直方图和局部二值模式特征的人体检测[J]. 光学精密工程, 2013,21(4): 1047-1053 DOI：

HUANG Yan FAN Ci-en ZHU Qiu-ping ZHANG Hu DENG De-xiang. HOG-LBP pedestrian detection[J]. Editorial Office of Optics and Precision Engineering, 2013,21(4): 1047-1053 DOI：

摘要

针对采用单一梯度方向直方图(HOG)特征进行人体检测时易受竖直梯度分量干扰的缺点，提出了将分块局部二值模式(LBP)特征加入HOG特征的方法。首先，将检测窗口划分为大小为1616的不重叠块，以块为单位统计LBP特征直方图，并通过大量实验获得了LBP算子的最佳参数；然后用优化过的插值方式计算HOG特征，将两者组成联合直方图。最后

用线性支持向量机(SVM)通过Bootstrapping的方式训练，得到判别模型。在INRIA人体库上的测试表明，检出率在误检率（FPPW）为10-4时由原始的89%提高到95%，单窗口检测速度由0.625 ms提高到0.533 ms。本文将纹理特征加入原始描述轮廓的HOG特征中，排除了部分梯度干扰信息造成的误检，提高了检出率。

Abstract

This paper proposed a method to concatenate a cell-structured Local Binary Pattern(LBP) feature into Histogram of Gradients(HOG) to solve the problem that HOG was vulnerable to the interference of vertical background gradient information in pedestrian detection. Firstly

the detection window was divided into 1616 non-overlapping blocks

then the LBP histogram of each block was calculated and his parameters were obtained by extensive experiments. Afterwards

the HOG was computed by the optimized interpolation method

and it was combined with LBP histogram to constitute a joint histogram. Finally

a discriminative model was trained by Bootstrapped linear Support Vector Machine(SVM). Based on the test of the INRIA pedestrian dataset

it is shown that the detection rate has been increased from 89% of the HOG feature to 95% when False Positive Per Window(FPPW) is 10-4

and the detection speed has been raised from 0.625 to 0.533 ms per window. It is concluded that the proposed method in this paper eliminates the false detection caused by the interference of gradient information and improves the detection rate by describing both contour and texture information.

关键词

Keywords

references

李欣，赵亦工，陈冰，等.基于模糊分类的弱小目标检测方法[J].光学精密工程.2009,17(9):2311-2320.LI X, ZHAO Y G, CHENG B, et al..Approach to dim and small target detection based on fuzzy classification [J]. Opt. Precision Eng., 2009, 17(9):2311-2320. (in Chinese)[2]刘亚洲. 基于时空分析和多粒度特征表示的人体检测方法研究[D].哈尔滨：哈尔滨工业大学，2009.LIU Y ZH. Spatial-temporal analysis and multi-granularity representation based human detection[D].Harbin: Harbin Institute of Technology,2009.(in Chinese)[3]LOWE D G.Distinctive image features from scale-invariant keypoints[J].International Journal of Computer Vision, 2004,60(2):91-110.[4]DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C].Computer Vision and Pattern Recognition, IEEE Computer Society Conference,2005:886-893.[5]WALK S, MAJER N, SCHINDLER K, et al..New features and insights for pedestrian detection[C].Computer Vision and Pattern Recognition,IEEE Computer Society Conference,2010:1030-1037.[6]WATANABE T, ITO S, YOKOI K.Co-occurrence histograms of oriented gradients for human detection [J].IPSJ Transactions on Computer Vision and Applications,2010,2:39-47.[7]MAJI S, BERG A C, MALIK J.Classification using intersection kernel support vector machines is efficient [C].Computer Vision and Pattern Recognition, IEEE Computer Society Conference,2008:1-8.[8]FELZENSZWALB P F, GIRSHICK R B, MCALLESTER D.Cascade object detection with deformable part models [C].Computer Vision and Pattern Recognition,IEEE Computer Society Conference,2010:2241-2248.[9]WANG X Y, HAN T X, YAN SH CH.An HOG-LBP human detector with partial occlusion handling[C]. Computer Vision, IEEE International Conference,2009:32-39.[10]ZHU Q, YEH M C, CHENG K T, et al..Fast human detection using a cascade of histograms of oriented gradients [C].Computer Vision and Pattern Recognition,IEEE Computer Society Conference, 2006:1491-1498.[11]PORIKLI F.Integral histogram: a fast way to extract histograms in Cartesian spaces [C].Computer Vision and Pattern Recognition,IEEE Computer Society Conference,2005:829-836. [12]BILGIC B.Fast human detection with cascaded ensembles [D].Massachusetts:Massachusetts Institute of Technology,2010.[13]DALAL N.Finding people in images and videos[D].France:the French National Institute for Research in Computer Science and Control,2006.[14]PANG Y W, YUAN Y, LI X L, et al.. Efficient HOG human detection [J]. Signal Processing, 2011, 91: 773-781.[15]GUO ZH H, ZHANG L,ZHANG D.A Completed modeling of local binary pattern operator for texture classification [J].IEEE Transactions on Image Processing,2010,19(6):1657-1663.[16]FAN R E,CHANG K W,HSIEH,et al.. LIBLINEAR:a library for large linear classification[J].Journal of Machine Learning Research,2008,9:1871-1874.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于改进SIFT特征提取的车标识别

采用自适应变异粒子群优化SVM的行为识别

结合形态学属性断面与支持向量机的合成孔径雷达图像变化检测

基于改进支持向量机的目标威胁估计

采用降维技术的红外目标检测与识别