• Laser & Optoelectronics Progress
  • Vol. 59, Issue 8, 0817003 (2022)
Liming Liang*, Jiang Yin, Yuanyuan Wu, and Jun Feng
Author Affiliations
  • School of Electrical Engineering and Automation, Jiangxi University of Science and Technology, Ganzhou , Jiangxi 341000, China
  • show less
    DOI: 10.3788/LOP202259.0817003 Cite this Article Set citation alerts
    Liming Liang, Jiang Yin, Yuanyuan Wu, Jun Feng. Medical Image Segmentation Algorithm Based on Bilateral Fusion[J]. Laser & Optoelectronics Progress, 2022, 59(8): 0817003 Copy Citation Text show less
    Overall structure of medical image segmentation network model BFNet based on bilateral fusion
    Fig. 1. Overall structure of medical image segmentation network model BFNet based on bilateral fusion
    Structure of deep semantic fusion network
    Fig. 2. Structure of deep semantic fusion network
    Structure of RFB
    Fig. 3. Structure of RFB
    Structure of DFB
    Fig. 4. Structure of DFB
    Structure of shallow detail fusion network
    Fig. 5. Structure of shallow detail fusion network
    Structure of partition head
    Fig. 6. Structure of partition head
    Comparison of visualization results of different algorithms on ISIC2017 dataset. (a) Original images; (b) manual labeling images by experts; (c) proposed algorithm; (d) Ref.[4]; (e) Ref.[7]; (f) Ref.[11]; (g) Ref.[27]
    Fig. 7. Comparison of visualization results of different algorithms on ISIC2017 dataset. (a) Original images; (b) manual labeling images by experts; (c) proposed algorithm; (d) Ref.[4]; (e) Ref.[7]; (f) Ref.[11]; (g) Ref.[27]
    AlgorithmmIoU /%mDice /%Precision /%Recall /%Overall accuracy /%
    Ref.[447.159.767.261.789.4
    Ref.[557.269.074.572.591.7
    Ref.[1061.371.478.474.291.7
    Ref.[373.783.188.283.595.2
    Ref.[1275.984.587.885.995.2
    Ref.[2373.381.386.184.094.9
    Ref.[2474.484.189.083.695.3
    Ref.[2577.685.789.186.296.1
    Ref.[2578.686.490.685.996.1
    Ref.[2681.087.694.486.096.8
    Ref.[2084.890.490.792.396.9
    BFNet (ours)86.292.389.186.796.1
    Table 1. Comparison results of proposed algorithm and other algorithms on Kvasir-SEG dataset
    AlgorithmmIoU /%Overall accuracy /%Kappa /%
    Ref.[489.1193.3287.34
    Ref.[789.1995.3388.46
    Ref.[1189.5395.4888.46
    Ref.[2789.6495.5089.96
    BFNet (ours)89.6695.5488.90
    Table 2. Comparison results of proposed algorithm and other algorithms on ISIC2016 dataset
    AlgorithmmIoU /%Overall accuracy /%Kappa /%
    Ref.[478.3491.5774.87
    Ref.[780.3292.4577.49
    Ref.[1180.1692.3877.28
    Ref.[2782.2393.2080.51
    BFNet (ours)83.4193.3981.45
    Table 3. Comparison results of proposed algorithm and other algorithms on ISIC2017 dataset
    Shallow networkDeep networkAggELmIoU /%mDice /%Overall accuracy /%
    BackboneRFBDFB
    79.8188.3092.00
    80.1188.5992.14
    83.0090.4893.11
    83.3490.6893.30
    82.6989.8893.00
    83.4190.7393.39
    Table 4. Influence of different modules on network performance
    Liming Liang, Jiang Yin, Yuanyuan Wu, Jun Feng. Medical Image Segmentation Algorithm Based on Bilateral Fusion[J]. Laser & Optoelectronics Progress, 2022, 59(8): 0817003
    Download Citation