Acta Optica Sinica, Volume. 45, Issue 15, 1510003(2025)

Difference Aware Guided Boundary Transformer Network for Childhood Pneumonia CT Image Segmentation

Jia Lü1,2、*, Mingkai Yu1, Xin Chen3, and Ling He3、**
Author Affiliations
  • 1College of Computer and Information Sciences, Chongqing Normal University, Chongqing 401331, China
  • 2National Center for Applied Mathematics in Chongqing, Chongqing Normal University, Chongqing 401331, China
  • 3Intelligent Application of Big Data in Pediatrics Engineering Research Center of Chongqing Education Commission of China, Ministry of Education Key Laboratory of Child Development and Disorders, National Clinical Research Center for Child Health and Disorders, Department of Radiology Children’s Hospital of Chongqing Medical University, Chongqing 400014, China
  • show less
    Figures & Tables(14)
    Structure of the DBTU-Net
    Structure of Gated Channel Transformation
    Structure of boundary Transformer module
    Sample examples in the Child-P dataset and their corresponding labels
    Confusion matrices for the four metrics under different hyperparameter configurations on the Child-P dataset. (a) JI; (b) SE; (c) MCC; (d) HD
    Segmentation results (up) and their confidence maps (down) on the Child-P dataset. (a) Label; (b) DBTU-Net; (c) U-Net; (d) U-Net++; (e) TransDeepLab; (f) CASCADE; (g) MEGANet
    Segmentation results (up) and their confidence maps (down) on the COVID and MosMed dataset. (a) Label; (b) DBTU-Net; (c) U-Net; (d) U-Net++; (e) TransDeepLab; (f) CASCADE; (g) MEGANet
    Local segmentation results and heatmaps of different methods on three datasets. (a) Original images; (b) local label; (c) proposed method; (d) U-Net; (e) CASCADE; (f) MEGANet
    Comparison of computational complexity on the Child-P dataset
    • Table 1. Evaluation metrics

      View table

      Table 1. Evaluation metrics

      MetricsDescription
      Dice /%2×|ŷy|/|ŷ|+|y|
      JI /%|ŷy|/|ŷy|
      SE /%NTP/NTP+NFN
      MCC /%NTP×NTN-NFP×NFNNTP+NFP×NTP+NFN×NTN+NFP×NTN+NFN1/2
      HD /pixelmaxmaxpPminyYp-y,maxyYminpPp-y
    • Table 2. Experimental results for different combinations of modules on the Child-P dataset

      View table

      Table 2. Experimental results for different combinations of modules on the Child-P dataset

      Combinations of modulesDice (↑)JI (↑)SE (↑)MCC (↑)HD /pixel (↓)
      U-Net (32)0.80480.67340.72770.806662.04
      U-Net (32)+GCT0.88020.78600.86660.878430.50
      U-Net (32)+DAF0.88690.79680.86710.885430.50
      U-Net (32)+BT0.88250.78970.86760.880833.17
      U-Net (32)+DAF+BT0.89120.80370.87960.889525.24
      U-Net (32)+GCT+DAF0.89200.80520.88560.890436.81
      U-Net (32)+GCT+BT0.88940.80080.88320.887630.00
      DBTU-Net0.89600.81170.89310.894425.08
    • Table 3. Comparison of experimental results with existing networks on the Child-P dataset

      View table

      Table 3. Comparison of experimental results with existing networks on the Child-P dataset

      NetworkDiceJISEMCCHD /pixel
      U-Net70.81440.68690.75470.814247.30
      ResU-Net80.80170.66900.73010.802756.82
      U-Net++90.83940.72330.79170.838546.32
      TMU-Net190.78930.65190.71450.790735.44
      TransDeepLab210.84660.73400.83170.844349.50
      CSU-Net220.87110.77160.87300.869058.01
      CASCADE310.88030.78620.87890.878354.64
      CHWS-UNet320.85250.74290.81530.851184.13
      MEGANet330.87460.77710.85020.873034.67
      DBTU-Net0.89610.81170.89310.894425.08
    • Table 4. Comparison of experimental results with existing networks on the COVID dataset

      View table

      Table 4. Comparison of experimental results with existing networks on the COVID dataset

      NetworkDiceJISEMCCHD /pixel
      U-Net70.77080.62710.71390.757422.56
      ResU-Net80.78460.64550.70770.775122.56
      U-Net++90.80850.67860.74290.798732.39
      TMU-Net190.78030.63970.69870.771428.20
      TransDeepLab210.77580.63370.71910.762635.52
      CSU-Net220.77340.63050.75010.756923.83
      CASCADE310.80520.67390.74210.794922.65
      CHWS-UNet320.77370.63090.73510.758428.64
      MEGANet330.79720.66290.72160.788333.62
      DBTU-Net0.82120.69660.77020.810421.93
    • Table 5. Comparison of experimental results with existing networks on the MosMed dataset

      View table

      Table 5. Comparison of experimental results with existing networks on the MosMed dataset

      NetworkDiceJISEMCCHD /pixel
      U-Net70.69830.53640.62050.701645.81
      ResU-Net80.69460.53210.59420.702741.68
      U-Net++90.71150.55220.62270.716842.65
      TMU-Net190.70430.54350.66300.703366.71
      TransDeepLab210.72380.56720.68660.722735.83
      CSU-Net220.65440.48630.56300.660836.29
      CASCADE310.73710.58360.66280.739824.29
      CHWS-UNet320.70080.53940.59160.711238.83
      MEGANet330.72050.56310.62810.726543.37
      DBTU-Net0.74370.59200.66880.746542.73
    Tools

    Get Citation

    Copy Citation Text

    Jia Lü, Mingkai Yu, Xin Chen, Ling He. Difference Aware Guided Boundary Transformer Network for Childhood Pneumonia CT Image Segmentation[J]. Acta Optica Sinica, 2025, 45(15): 1510003

    Download Citation

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

    Category: Image Processing

    Received: Mar. 18, 2025

    Accepted: May. 6, 2025

    Published Online: Aug. 8, 2025

    The Author Email: Jia Lü (lvjia@cqnu.edu.cn), Ling He (heling508@sina.com)

    DOI:10.3788/AOS250760

    CSTR:32393.14.AOS250760

    Topics