Optics and Precision Engineering, Volume. 31, Issue 6, 936(2023)

Bone scintigraphic classification method based on ACGAN and transfer learning

Hong YU1,*... Renze LUO1, Chunmeng CHEN2, Xiang TANG3 and Renquan LUO1 |Show fewer author(s)
Author Affiliations
  • 1College of Electrical Engineering and Information,Southwest Petroleum University, Chengdu60500, China
  • 2Department of Nuclear Medicine, The No. People’s Hospital of Yibin, Yibin644000, China
  • 3College of Computer Science, Southwest Petroleum University, Chengdu610500, China
  • show less
    Figures & Tables(22)
    ACGAN principle flow
    Overall flow chart of the method in this paper
    Improved network models
    CA modules
    Residual gated attention architecture
    Dense residual attention module
    CNN classification model
    Comparison of MU-ACGAN training process
    Comparison of transfer learning
    Bone images generated by MU-ACGAN
    Generative bone imaging and real bone imaging
    Similarity comparison
    Bone scintigraphy to identify heatmaps
    Confusion matrix
    • Table 1. Number and ratio of training and testing of bone imaging

      View table
      View in Article

      Table 1. Number and ratio of training and testing of bone imaging

      IndexHealthMetastasesBenign changes
      Train/TestTrain/TestTrain/Test
      Quantity130/32101/2543/15
      Total16212658
      Ratio46.82%36.42%16.76%
    • Table 2. Amount of bone imaging data after the expansion

      View table
      View in Article

      Table 2. Amount of bone imaging data after the expansion

      ClassHealthMetastasesBenign changes
      Quantity1 9501 515645
    • Table 3. Quality assessment of bone imaging

      View table
      View in Article

      Table 3. Quality assessment of bone imaging

      Image typeHealthMetastasesBenign changes
      ACGAN0.458 50.516 00.506 8
      MU-ACGAN0.539 00.557 20.548 5
    • Table 4. Classification effects of different data augmentation methods

      View table
      View in Article

      Table 4. Classification effects of different data augmentation methods

      NumberAccuracyPrecisionRecallSpecificityF1-socre
      10.701 30.695 30.665 70.680 10.841 6
      20.727 30.730 10.687 80.708 30.845 2
      30.740 30.737 50.675 20.705 00.850 8
      40.779 20.803 20.692 80.743 90.868 8
      50.805 20.866 20.728 40.791 30.882 5
    • Table 5. Comparison of transfer learning strategies

      View table
      View in Article

      Table 5. Comparison of transfer learning strategies

      TypeAccuracyRecall
      Transfer learning0.805 20.728 4
      Dataset50.662 30.589 9
      Training directly0.597 40.527 6
    • Table 6. Experimental evaluation of multi-scale CNN model ablation

      View table
      View in Article

      Table 6. Experimental evaluation of multi-scale CNN model ablation

      ModelsAccuracyRecallF1-socre
      VGG160.766 20.666 30.743 6
      DenseNet1210.805 20.755 00.767 7
      ResNext500.766 20.696 80.736 7
      Multi-scale CNN0.831 20.763 90.817 8
    • Table 7. Comparison of the experimental results of different CNN models

      View table
      View in Article

      Table 7. Comparison of the experimental results of different CNN models

      ModelsAccuracyPrecisionRecallSpecificityF1-socre
      ResNet340.779 20.803 20.692 80.868 80.743 9
      MobileNetV30.779 20.785 70.745 90.875 60.765 2
      InceptionV30.792 20.847 70.750 60.874 20.796 2
      EfficientNet0.805 20.868 40.741 80.880 60.800 1
      Ours0.831 20.879 90.763 90.899 20.817 8
    • Table 8. Evaluation metrics of different CNN classification models

      View table
      View in Article

      Table 8. Evaluation metrics of different CNN classification models

      DatasetAccuracyRecallSpecificity
      Dataset40.831 20.763 90.899 2
      Dataset50.857 10.804 00.912 0
    Tools

    Get Citation

    Copy Citation Text

    Hong YU, Renze LUO, Chunmeng CHEN, Xiang TANG, Renquan LUO. Bone scintigraphic classification method based on ACGAN and transfer learning[J]. Optics and Precision Engineering, 2023, 31(6): 936

    Download Citation

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

    Category: Information Sciences

    Received: Jul. 19, 2022

    Accepted: --

    Published Online: Apr. 4, 2023

    The Author Email: YU Hong (790622472@qq.com)

    DOI:10.37188/OPE.20233106.0936

    Topics