Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 57, Issue 24, 241503(2020)

Improved Algorithm for Person Re-Identification Based on Global Features

Tao Zhang, Zhengming Yi*, Xuan Li, and Xing Sun
Author Affiliations
  • School of Electrical and Information Engineering, Tianjin University, Tianjin 300072,China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (7)
    Equations (0)
    References (26)
    Cited By (7)
    Tools
    Paper Information
    Topics
    Figures & Tables(7)
    Learning process of triplet loss fuction

    Fig. 1. Learning process of triplet loss fuction

    Download full size
    Structure of network model

    Fig. 2. Structure of network model

    Download full size
    Performance curves for different λ values on different datasets. (a) Market1501 dataset; (b) DukeMTMC-reID dataset

    Fig. 3. Performance curves for different λ values on different datasets. (a) Market1501 dataset; (b) DukeMTMC-reID dataset

    Download full size
    Comparative experimental results of different algorithms on different datasets. (a) Market1501 dataset; (b) DukeMTMC-reID dataset

    Fig. 4. Comparative experimental results of different algorithms on different datasets. (a) Market1501 dataset; (b) DukeMTMC-reID dataset

    Download full size

    • Table 1. Performance comparison of different parameters

      View table

      Table 1. Performance comparison of different parameters

      λMarket1501DukeMTMC-reID
      mAP/%rank1/%mAP/%rank1/%
      0.187.494.878.288.0
      0.387.594.778.488.2
      0.587.194.777.887.7
      0.786.894.777.387.8
      0.986.494.676.787.6
      1.086.394.776.287.5

    • Table 2. Performance comparison of proposed algorithm with other advanced algorithms on Market1501 dataset and DukeMTMC-reID dataset

      View table

      Table 2. Performance comparison of proposed algorithm with other advanced algorithms on Market1501 dataset and DukeMTMC-reID dataset

      AlgorithmMarket1501DukeMTMC-reID
      mAP/%rank1/%mAP/%rank1/%
      AlignedReID++[15]77.790.667.481.2
      SCPNet[16]75.291.262.680.3
      PCB+RPP[17]81.693.869.283.3
      Pyramid[18]88.295.779.089.9
      HPM[19]82.794.274.386.6
      MGN[20]86.995.778.488.7
      IDE[21]59.979.5--
      SVDNet[22]62.182.356.876.7
      TriNet[9]69.184.9--
      AWTL[23]75.789.563.479.8
      Baseline[7]85.994.576.486.4
      Proposed algorithm87.594.778.488.2

    • Table 3. Performance comparison of proposed algorithm with other advanced algorithms on CUHK03 dataset

      View table

      Table 3. Performance comparison of proposed algorithm with other advanced algorithms on CUHK03 dataset

      AlgorithmCUHK03-LabeledCUHK03-Detected
      mAP/%rank1/%mAP/%rank1/%
      MGN[20]67.468.066.068.0
      PCB+RPP[17]--57.563.7
      MLFN[24]49.254.747.852.8
      HA-CNN[25]41.044.438.641.7
      SVDNet[22]37.840.937.341.5
      PAN[26]35.036.934.036.3
      IDE[21]21.022.219.721.3
      Proposed algorithm67.369.963.865.4
    Tools

    Get Citation

    Copy Citation Text

    Tao Zhang, Zhengming Yi, Xuan Li, Xing Sun. Improved Algorithm for Person Re-Identification Based on Global Features[J]. Laser & Optoelectronics Progress, 2020, 57(24): 241503

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Category: Machine Vision

    Received: May. 6, 2020

    Accepted: Jun. 9, 2020

    Published Online: Dec. 1, 2020

    The Author Email: Yi Zhengming (yzm411522@163.com)

    DOI:10.3788/LOP57.241503

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology