Optics and Precision Engineering, Volume. 31, Issue 17, 2555(2023)

Partial optimal transport-based domain adaptation for hyperspectral image classification

Bilin WANG1... Shengsheng WANG1,* and Zhe ZHANG2 |Show fewer author(s)
Author Affiliations
  • 1College of Computer Science and Technology, Jilin University, Changchun3002, China
  • 2Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Sciences, Suzhou15163, China
  • show less
    Figures & Tables(8)
    Framework of proposed domain adaptive model base on partial optimal transport
    Pseudocolor images and corresponding ground truths with color indexes of Pavia dataset
    Pseudocolor images and corresponding ground truths with color indexes of Houston dataset
    Classification maps of PC dataset produced by different methods
    • Table 1. Land cover classes and numbers of samples in Pavia dataset

      View table
      View in Article

      Table 1. Land cover classes and numbers of samples in Pavia dataset

      类别样本数量
      UP(源域)PC(目标域)
      牧场18 6493 090
      阴影9472 863
      砖块3 6822 685
      沥青道路6 6319 248
      树木3 0647 598
      沥青1 3307 287
      裸露土壤5 0296 584
      总计39 33239 355
    • Table 2. Land cover classes and numbers of samples in Houston dataset

      View table
      View in Article

      Table 2. Land cover classes and numbers of samples in Houston dataset

      类别样本数量
      H13(源域)H18(目标域)
      草地3451 353
      道路4436 365
      受压草地3652 766
      水面28522
      树木3652 766
      居民建筑3195 347
      非居民建筑40832 459
      总计2 53053 200
    • Table 3. Classification performance on OA,AA(%) and Kappa of different target domains

      View table
      View in Article

      Table 3. Classification performance on OA,AA(%) and Kappa of different target domains

      MethodUP→PCH13→H18
      OA(%)AA(%)K(×100)OA(%)AA(%)K(×100)
      TSVM76.03±0.0272.41±2.2270.70±3.0552.03±0.3262.40±0.4034.29±0.80
      DAN80.68±0.0276.18±1.6276.47±2.3461.45±2.1873.47±1.7447.08±2.36
      DANN83.90±0.0578.40±0.4780.41±0.5062.61±2.1065.09±3.5449.33±1.98
      ED-DMM-UDA82.50±0.5077.80±0.5678.73±0.5466.09±1.8374.48±1.6653.10±2.18
      CDA89.78±3.0388.80±5.0587.71±3.6668.44±0.7775.31±1.5855.65±0.77
      CLDA92.80±0.6791.83±1.0991.32±0.8170.12±2.4379.07±0.3857.75±2.49
      Ours93.67±0.5094.17±1.0693.76±0.7170.25±2.2079.89±1.0658.32±1.22
    • Table 4. Ablation study classification results of OA with different transfer tasks

      View table
      View in Article

      Table 4. Ablation study classification results of OA with different transfer tasks

      MethodUP→PCH13→H18
      baseline91.4667.50
      +OT91.9668.24
      +OT+CS92.8569.27
      +POT(s=0.5)+CS93.0569.43
      +POT(s-adapt)+CS93.6770.25
    Tools

    Get Citation

    Copy Citation Text

    Bilin WANG, Shengsheng WANG, Zhe ZHANG. Partial optimal transport-based domain adaptation for hyperspectral image classification[J]. Optics and Precision Engineering, 2023, 31(17): 2555

    Download Citation

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

    Category: Information Sciences

    Received: Feb. 15, 2023

    Accepted: --

    Published Online: Oct. 9, 2023

    The Author Email: WANG Shengsheng (wss@jlu.edu.cn)

    DOI:10.37188/OPE.20233117.2555

    Topics