Laser & Optoelectronics Progress, Volume. 58, Issue 16, 1610022(2021)

Gender Classification of Iris Image Based on Residual Network

Fusheng Yu1,2,3, Jiang Yu1、*, Yuanfu Lu2,3、**, Zhisheng Zhou2,3, and Guangyuan Li2,3
Author Affiliations
  • 1School of Information, Yunnan University, Kunming, Yunnan 650000, China
  • 2Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong 518000, China
  • 3Key Laboratory of Human-Machine Intelligence-Synergy Systems, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, Guangdong 518055, China
  • show less
    Figures & Tables(11)
    Process flowchart of ResNet model
    Structure of ResNet50
    Structure of residual element
    Effect comparison between transfer learning and non-transfer learning
    Different gender iris images. (a) Women; (b) male
    Dataset expansion images. (a) 90° left rotation; (b) 90° right rotation; (c) flip upside down
    Images of different processes. (a) Positioning process; (b) standardization process; (c) integration process
    Performance curves under different parameters. (a) Loss rate; (b) accuracy
    Accuracy comparison curves of three networks
    • Table 1. Partition of experimental data

      View table

      Table 1. Partition of experimental data

      Dataset typeProportion of data /%Data
      Training dataset7011200
      Validation dataset203200
      Test dataset101600
    • Table 2. Experimental results under different parameters

      View table

      Table 2. Experimental results under different parameters

      EpochBatch sizeOptimizerTime /sACC /%
      2064SGD31087.56
      2064Adam33489.12
      2064Adagrad40688.36
      20128SGD18088.61
      20128Adam20991.53
      20128Adagrad22888.43
      20256SGD10389.21
      20256Adam12791.40
      20256Adagrad13989.13
      4064SGD60589.23
      EpochBatch sizeOptimizerTime /sACC /%
      4064Adam70191.91
      4064Adagrad82390.34
      40128SGD34692.62
      40128Adam38494.60
      40128Adagrad42093.38
      40256SGD18291.57
      40256Adam20893.36
      40256Adagrad23493.80
      6064SGD85089.35
      6064Adam94391.86
      6064Adagrad110090.68
      60128SGD43293.12
      60128Adam51094.30
      60128Adagrad62393.70
      60256SGD22593.83
      60256Adam27493.62
      60256Adagrad32893.70
    Tools

    Get Citation

    Copy Citation Text

    Fusheng Yu, Jiang Yu, Yuanfu Lu, Zhisheng Zhou, Guangyuan Li. Gender Classification of Iris Image Based on Residual Network[J]. Laser & Optoelectronics Progress, 2021, 58(16): 1610022

    Download Citation

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

    Category: Image Processing

    Received: Aug. 5, 2020

    Accepted: Oct. 10, 2020

    Published Online: Aug. 16, 2021

    The Author Email: Jiang Yu (yujiang@ynu.edu.cn), Yuanfu Lu (yf.lu@siat.ac.cn)

    DOI:10.3788/LOP202158.1610022

    Topics