• Acta Photonica Sinica
  • Vol. 53, Issue 1, 0111003 (2024)
Hong HUANG1,*, Yichuan YANG1, Long WANG1, Fujian ZHENG1, and Jian WU2
Author Affiliations
  • 1Key Laboratory of Optoelectronic Technology and System,Ministry of Education,Chongqing University,Chongqing 400044,China
  • 2Head and Neck Cancer Centre,Chongqing University Cancer Hospital & Chongqing Cancer Institute & Chongqing Cancer Hospital,Chongqing 400030,China
  • show less
    DOI: 10.3788/gzxb20245301.0111003 Cite this Article
    Hong HUANG, Yichuan YANG, Long WANG, Fujian ZHENG, Jian WU. Boundary Perception Network for Pathological Image Segmentation[J]. Acta Photonica Sinica, 2024, 53(1): 0111003 Copy Citation Text show less
    The overall structure of the BPNet algorithm
    Fig. 1. The overall structure of the BPNet algorithm
    The structure of attention boundary perception module
    Fig. 2. The structure of attention boundary perception module
    The structure of adaptive shuffle channel attention module
    Fig. 3. The structure of adaptive shuffle channel attention module
    The GlaS pathological image datasets
    Fig. 4. The GlaS pathological image datasets
    The MoNuSeg nuclei datasets
    Fig. 5. The MoNuSeg nuclei datasets
    The segmentation results of different algorithms on GlaS datasets
    Fig. 6. The segmentation results of different algorithms on GlaS datasets
    The segmentation results of ablation experiments on GlaS datasets
    Fig. 7. The segmentation results of ablation experiments on GlaS datasets
    The segmentation results of different algorithms on MoNuSeg datasets
    Fig. 8. The segmentation results of different algorithms on MoNuSeg datasets
    AlgorithmDice/%IoU/%ACC/%PRE/%Parameters/(×106
    U-Net(2015)83.40±3.2671.84±5.8683.04±3.8379.99±6.5131.0
    UNet++(2018)85.11±0.7374.10±1.1085.48±0.4185.83±0.9739.4
    AttentionUNet(2018)86.30±2.0875.99±4.0086.39±2.9483.99±5.8263.1
    MultiResUNet(2020)87.20±0.9977.38±1.5586.98±0.6986.29±0.8559.1
    MedT(2021)82.86±0.9770.76±1.4182.46±0.7883.93±0.997.0
    TransUNet(2021)88.67±0.6579.66±1.0688.46±0.6990.00±0.78421.2
    UCTransNet(2022)89.39±0.6880.83±1.1187.93±0.9387.00±1.6565.5
    BPNet92.21±0.1985.55±0.3392.14±0.1592.07±0.9364.9
    Table 1. The experimental results with different methods on GlaS datasets(Mean ± Std)
    AlgorithmDice/%IoU/%ACC/%PRE/%Parameters/(×106
    Baseline87.09±1.9778.10±3.3388.65±2.1688.54±3.1063.3
    Baseline+BPM91.79±0.0984.81±0.1691.80±0.0792.87±0.3064.8
    Baseline+BPM+ASCAM92.21±0.1985.55±0.3392.14±0.1592.07±0.9364.9
    Table 2. The ablation experimental results on GlaS datasets(Mean ± Std)
    AlgorithmDice/%IoU/%ACC/%PRE/%Parameters/(×106
    U-Net(2015)77.17±2.8462.90±3.7989.90±1.6265.01±3.6831.0
    UNet++(2018)78.90±0.5765.16±0.7890.98±0.3367.42±1.3239.4
    AttentionUNet(2018)75.00±1.8660.03±2.3888.40±1.4260.64±3.3963.1
    MultiResUNet(2020)79.70±0.6166.26±0.8491.52±0.3369.19±1.6159.1
    MedT(2021)74.48±0.6159.34±0.7789.15±0.1663.61±3.307.0
    TransUNet(2021)80.50±0.3967.37±0.5491.40±0.2473.42±1.48421.2
    UCTransNet(2022)79.08±0.7965.58±1.0091.02±0.5370.38±1.1365.5
    BPNet81.18±0.4468.34±0.6592.50±0.2475.46±2.1564.9
    Table 3. The experimental results with different methods on MoNuSeg datasets(Mean ± Std)
    AlgorithmDice/%IoU/%ACC/%PRE/%Parameters/(×106
    Baseline77.74±0.7363.59±0.9891.47±0.0673.38±2.7463.3
    Baseline+BPM79.96±0.4666.61±0.6492.02±0.3673.97±4.2164.8
    Baseline+BPM+ASCAM81.18±0.4468.34±0.6592.50±0.2475.46±2.1564.9
    Table 4. The ablation experimental results on MoNuSeg datasets(Mean ± Std)
    Hong HUANG, Yichuan YANG, Long WANG, Fujian ZHENG, Jian WU. Boundary Perception Network for Pathological Image Segmentation[J]. Acta Photonica Sinica, 2024, 53(1): 0111003
    Download Citation