Table 6 Performance analysis of the proposed ACU-Net model with existing works.
From: Enhancing brain tumor segmentation using attention based convolutional UNet on MRI images
Model | Dataset | Dice WT | Dice TC | Dice ET | HD95 WT | HD95 TC | HD95 ET | ASSD WT | ASSD TC | ASSD ET |
|---|---|---|---|---|---|---|---|---|---|---|
3D-UNet4 | BRATS 2018 | 91.17 | 84.11 | 77.00 | 6.02 | 7.25 | 8.40 | 2.30 | 2.85 | 3.40 |
HTTU-Net5 | BRATS 2018 | 91.50 | 92.30 | 88.70 | 5.80 | 6.90 | 7.85 | 2.10 | 2.50 | 3.10 |
RMU-Net31 | BRATS 2018 | 90.80 | 86.75 | 79.36 | 6.10 | 7.10 | 8.10 | 2.45 | 2.90 | 3.50 |
CNN10 | BRATS 2018 | 89.93 | 92.11 | 92.23 | 6.45 | 6.85 | 7.95 | 2.50 | 2.75 | 3.30 |
CNN11 | BRATS 2018 | 91.20 | 88.34 | 81.84 | 5.92 | 7.05 | 8.20 | 2.15 | 2.80 | 3.20 |
3D-UNet12 | BRATS 2018 | 90.00 | 83.00 | 71.00 | 6.78 | 7.50 | 8.70 | 2.60 | 3.00 | 3.60 |
ACU-Net (Proposed) | BRATS 2018 | 94.04 | 98.63 | 98.77 | 3.50 | 3.20 | 3.10 | 1.20 | 1.10 | 1.05 |
