留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

融合底层和中层字典特征的行人重识别

王丽

王丽. 融合底层和中层字典特征的行人重识别[J]. 中国光学(中英文), 2016, 9(5): 540-546. doi: 10.3788/CO.20160905.0540
引用本文: 王丽. 融合底层和中层字典特征的行人重识别[J]. 中国光学(中英文), 2016, 9(5): 540-546. doi: 10.3788/CO.20160905.0540
WANG Li. Pedestrian re-identification based on fusing low-level and mid-level features[J]. Chinese Optics, 2016, 9(5): 540-546. doi: 10.3788/CO.20160905.0540
Citation: WANG Li. Pedestrian re-identification based on fusing low-level and mid-level features[J]. Chinese Optics, 2016, 9(5): 540-546. doi: 10.3788/CO.20160905.0540

融合底层和中层字典特征的行人重识别

doi: 10.3788/CO.20160905.0540
详细信息
    通讯作者:

    王丽(1979-), 女, 吉林长春人, 学士, 工程师, 主要从事信息通信技术方面的研究.E-mail:44417020@qq.com

  • 中图分类号: TP394.1

Pedestrian re-identification based on fusing low-level and mid-level features

More Information
  • 摘要: 针对当前行人重识别方法采用单一底层特征识别率较低的问题,提出一种融合底层和中层特征的识别方法,由粗到精对人体目标进行匹配识别。首先提取目标的颜色直方图和纹理直方图进行粗分类;然后将人体目标分为头部、躯干和腿部3个部分。忽略包含信息量较少的头部,对躯干和腿部,提出一种中层图像块字典提取方法,并对照该字典生成中层特征,进行精确分类。底层特征结合中层特征使算法既具有较好的区分度,又具有良好的泛化能力。实验结果表明本文算法在VIPeR数据库上的nAUC比已有方法提高6.3%,对遮挡和背景粘连的鲁棒性更好。

     

  • 图 1  行人不同部位划分

    Figure 1.  Segmentation of different body parts

    图 2  VIPeR图库上匹配结果

    Figure 2.  Matching result in VIPeR database

    图 3  ETHZ数据库上多帧匹配结果

    Figure 3.  Multi-frame matching rate in ETHZ database

    表  1  算法排名等级和nAUC对比

    Table  1.   Comparation of ranking matching rate and nAUC

    MethodRank-1Rank-10Rank-20Rank-30nAUC
    SDALF21.851.465.776.883.5
    ELF19.445.660.570.979.6
    SCEAF24.657.370.481.585.4
    Proposed method37.870.977.286.791.7
    下载: 导出CSV
  • [1] GONG S, CRISTANI M, YAN S, et al..Person Re-identification[M].London:Springer, 2014.
    [2] WANG X, DORETTO G, SEBASTIAN T B, et al..Shape and appearance context modeling[J].IEEE, 2007, 1(1):1-8. http://www.docin.com/p-1479988732.html
    [3] ARENZENA M, BAZZANI L, PERINA A, et al..Person re-identification by symmetry-driven accumulation of local features[C].IEEE Conference on Computer Vision and Pattern Recognition, San Francisco, USA, 2010:2360-2367.
    [4] CHENG D, CRISTANI M, STOPPA M, et al..Custom pictorial structures for re-identification[C].British Machine Vision Conference, Dundee, UK, 2011:749-760.
    [5] KOSTINGER M, HIRZER M, WOHLHART P, et al..Large scale metric learning from equivalence constraints[C].IEEE Conference on Computer Vision and Pattern Recognition, IEEE, 2012:2288-2295.
    [6] MA B, SU Y, JURIE F.Local descriptors encoded by fisher vectors for person re-identification[C].European Conference on Computer Vision, Florence, Italy, 2012:413-422.
    [7] ZHENG W, GONG S, XIANG T.Re-identification by Relative Distance Comparison[J].IEEE, 2013, 35(3):653-668.
    [8] 王睿, 朱正丹.融合全局-颜色信息的尺度不变特征变换[J].光学精密工程, 2015, 23(1): 295-301. doi: 10.3788/OPE.

    WANG R, ZHU ZH D.SIFT matching with color invariant characteristics and global context[J].Opt.Precision Eng., 2015, 23(1):295-301.(in Chinese) doi: 10.3788/OPE.
    [9] 王飞宇, 邸男, 贾平.结合尺度空间FAST角点检测器和SURF描绘器的图像特征[J].液晶与显示, 2014, 29(4):598-604. doi: 10.3788/YJYXS

    WANG F Y, DI N, JIA P.Image features using scale-space FAST corner detector and SURF descriptor[J].Chinese J.Liquid Crystals and Displays, 2014, 29(4):598-604.(in Chinese) doi: 10.3788/YJYXS
    [10] 王晓华, 孙小姣.联合Gabor降维特征与奇异值特征的人脸识别[J].光学精密工程, 2015, 23(10):553-558. http://cpfd.cnki.com.cn/Article/CPFDTOTAL-GXJM201507001084.htm

    WANG X H, SUN X J.Face recognition based on Gabor reduction dimensionality features and singular value decomposition features[J].Opt.Precision Eng., 2015, 23(10):553-558.(in Chinese) http://cpfd.cnki.com.cn/Article/CPFDTOTAL-GXJM201507001084.htm
    [11] 邓丹, 吴谨, 朱磊, 等.基于纹理抑制和连续分布估计的显著性目标检测方法[J].液晶与显示, 2015, 30(1):120-125. doi: 10.3788/YJYXS

    DENG D, WU J, ZHU L, et al.Significant target detection method based on texture inhibition and continuous distribution estimation[J].Chinese J.Liquid Crystals and Displays, 2015, 30(1):120-125.(in Chinese) doi: 10.3788/YJYXS
    [12] BIRCHFIELD S T, RANGARAJAN S.Spatiograms versus histograms for region-based tracking[J].IEEE, 2005(2):1158-1163. http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.73.3012
    [13] SINGH S, GUPTA A, EFROS A A.Unsupervised discovery of mid-level discriminative patches[C].European Conference on Computer Vision, Florence, Italy, 2012:73-86.
    [14] JAIN A, GUPTA A, RODRIGUEZ M, et al..Representing videos using mid-level discriminative patches[C].IEEE Conference on Computer Vision and Pattern Recognition, IEEE, 2013:571-2578.
    [15] 陈莹, 朱明, 刘剑, 等.高斯混合模型自适应微光图像增强[J].液晶与显示, 2015, 30(2):300-309. doi: 10.3788/YJYXS

    CHEN Y, ZHU M, LIU J, et al..Automatic low light level image enhancement using Gaussian mixture modeling[J].Chinese J.Liquid Crystals and Displays, 2015, 30(2):300-309.(in Chinese) doi: 10.3788/YJYXS
    [16] GRAY D, TAO H.Viewpoint invariant pedestrian recognition with an ensemble of localized features[C].European Conference on Computer Vision, Florence, Marseille, Italy, 2008:262-275.
    [17] HU Y, LIAO S, LEI Z, et al..Exploring structural information and fusing multiple features for person re-identification[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Portland, USA, 2013:794-799.
    [18] SCHWARTZ W, DAVIS L.Learning discriminative appearance based models using partial least squares[C].Computer Graphics and Image Processing, Rio de Janeiro, Brazil, 2009:322-329.
  • 加载中
图(3) / 表(1)
计量
  • 文章访问数:  1217
  • HTML全文浏览量:  264
  • PDF下载量:  857
  • 被引次数: 0
出版历程
  • 收稿日期:  2016-04-05
  • 修回日期:  2016-05-26
  • 刊出日期:  2016-10-01

目录

    /

    返回文章
    返回