Chinese Journal of Liquid Crystals and Displays, Volume. 37, Issue 10, 1364(2022)

Improved generative adversarial network for pedestrian re-identification in clothes

Yu-xia ZHANG1,2, Jin CHE1,2、*, and Yu-ting HE1,2
Author Affiliations
  • 1School of Physics and Electronic-Electrical Engineering,Ningxia University,Yinchuan 750021,China
  • 2Ningxia Key Laboratory of Intelligent Sensing for Desert Information,Yinchuan 750021,China
  • show less
    Figures & Tables(13)
    Network structure
    Generate network frame diagram.(a)Self generation network;(b)Mutual generation network.
    Generator block diagram
    DenseBlock structure diagram
    Discriminator network frame diagram
    Discriminator network structure diagram
    Channel attention mechanism NAM
    Examples of images generated by swapping appearance and structure
    Examples of images generated by switching the appearance or structure code
    A few failure cases
    • Table 1. Public datasets details

      View table
      View in Article

      Table 1. Public datasets details

      数据集IDTrain IDTrain imageTest IDTest imageBbox/IDCam
      Market1501[21]1 50175112 93675019 73221.86
      DukeMTMC-reID[22]1 40470216 52270217 66125.98
    • Table 2. Comparison of the algorithm in this paper with existing algorithms

      View table
      View in Article

      Table 2. Comparison of the algorithm in this paper with existing algorithms

      ModelMarket1501DukeMTMC-reID
      Rank-1/%mAP/%Rank-1/%mAP/%
      PTGAN1287.768.978.556.9
      PNGAN1489.472.673.653.2
      FDGAN1590.577.480.064.5
      DG-Net1794.886.086.674.8
      CSL+MGL2395.386.188.176.4
      PAGAN2494.185.284.774.2
      AP-GAN1393.482.284.370.5
      TransReID2595.289.590.782.6
      Euc2693.779.788.376.8
      FGSAM2791.585.485.974.1
      HOReID2894.285.186.975.6
      MOPN2994.685.188.076.2
      PAFM3095.188.290.678.6
      Ours95.688.687.175.7
    • Table 3. Performance evaluation of different improvement points

      View table
      View in Article

      Table 3. Performance evaluation of different improvement points

      不同改进点Market1501DukeMTMC-reID
      (1)(2)Rank-1/%mAP/%Rank-1/%mAP/%
      ××94.886.086.674.8
      ×95.087.686.975.3
      ×95.287.986.875.0
      95.688.687.175.7
    Tools

    Get Citation

    Copy Citation Text

    Yu-xia ZHANG, Jin CHE, Yu-ting HE. Improved generative adversarial network for pedestrian re-identification in clothes[J]. Chinese Journal of Liquid Crystals and Displays, 2022, 37(10): 1364

    Download Citation

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

    Category: Research Articles

    Received: Mar. 15, 2022

    Accepted: --

    Published Online: Oct. 10, 2022

    The Author Email: Jin CHE (koalache@126.com)

    DOI:10.37188/CJLCD.2022-0086

    Topics