Table 1 Comparisons with the state-of-the-art baselines on the Kvasir-SEG dataset terms
Method | Year | mDice | mIoU | Recall | Precision |
|---|---|---|---|---|---|
ResUNet [42] | 2018 | 0.7907 | 0.4287 | 0.6909 | 0.8713 |
ResUNet++ [43] | 2019 | 0.8133 | 0.7927 | 0.8774 | 0.7064 |
U-Net [5] | 2015 | 0.8180 | 0.7460 | 0.6306 | 0.9222 |
U-Net++ [7] | 2018 | 0.8210 | 0.7430 | – | – |
HRNetV2-W48 [44] | 2020 | 0.8896 | 0.8262 | 0.8973 | 0.9056 |
DS-TransUNet-B [14] | 2021 | 0.9110 | 0.8561 | 0.9352 | 0.9143 |
DS-TransUNet-L [14] | 2021 | 0.9130 | 0.8592 | 0.9360 | 0.9164 |
TransFuse [30] | 2021 | 0.9180 | 0.8680 | – | – |
MSRF-Net [41] | 2021 | 0.9217 | 0.8914 | 0.9198 | 0.9666 |
EG-TransUNet (ours) | – | 0.9344 | 0.8927 | 0.9401 | 0.9436 |