Comparative prediction results on the ISIC2017 dataset
Dataset | Models | Year | DSC↑ | SE↑ | SP↑ | ACC↑ |
---|---|---|---|---|---|---|
ISIC2017 | U-Net* | 2015 | 0.8989 | 0.8793 | 0.9812 | 0.9613 |
SCR-Net* | 2021 | 0.8898 | 0.8497 | 0.9853 | 0.9588 | |
C2SDG* | 2021 | 0.8938 | 0.8859 | 0.9765 | 0.9588 | |
ATTENTION SWIN U-NET* | 2022 | 0.8859 | 0.8492 | 0.9847 | 0.9591 | |
MALUNet* | 2022 | 0.8896 | 0.8824 | 0.9762 | 0.9583 | |
UNeXt-S# | 2022 | 0.9017 | 0.8894 | 0.9806 | 0.9633 | |
EGE-UNet* | 2023 | 0.9073 | 0.8931 | 0.9816 | 0.9642 | |
VM-UNet* | 2024 | 0.9070 | 0.8837 | 0.9842 | 0.9645 | |
VM-UNet v2* | 2024 | 0.9045 | 0.8768 | 0.9849 | 0.9637 | |
LightM-UNet* | 2024 | 0.9080 | 0.8839 | 0.9846 | 0.9649 | |
UltraLight VM-UNet* | 2024 | 0.9091 | 0.9053 | 0.9790 | 0.9646 | |
UltraLight VM-UNet# | 2024 | 0.9097 | 0.9042 | 0.9804 | 0.9660 | |
UCM-Net (Baseline) | 2024 | 0.9120 | 0.8824 | 0.9877 | 0.9678 | |
MUCM-Net (1-patch) | 2024 | 0.9160 | 0.9090 | 0.9869 | 0.9689 | |
MUCM-Net (2-patch) | 2024 | 0.9126 | 0.9008 | 0.9829 | 0.9679 | |
MUCM-Net (4-patch) | 2024 | 0.9133 | 0.8871 | 0.9870 | 0.9681 | |
MUCM-Net (8-patch) | 2024 | 0.9185 | 0.9014 | 0.9857 | 0.9697 |
* the results cited from UltraLight VM-Net; # the results tested by us. ISIC: International Skin Imaging Collaboration; DSC: Dice Similarity Coefficient; SE: sensitivity; SP: specificity; ACC: accuracy. ↑: the higher the value, the better the performance