Lesion Change Detection on Longitudinal Multiple Sclerosis Brain Imaging dataset (test)
75.8LTPRSiamese U-net
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Siamese U-netLoss=FTL, gamma=1, L2 regularization=10^-6, Blob size=202021.06 | 75.8 | 94.2 | 5.79 | |
| Siamese U-netLoss=FTL, gamma=0.75, L2 regularization=10^-6, Blob size=202021.06 | 74.7 | 94.6 | 5.41 | |
| Siamese U-netLoss=FTL, gamma=1, L2 regularization=10^-4, Blob size=202021.06 | 70.8 | 88.4 | 11.6 | |
| U-net-AG-MS-DSLoss=FTL, gamma=1, Attention Gate=true, Multiscale Input=true, Deep Supervision=true, Blob size=202021.06 | 66.4 | 94.5 | 5.46 | |
| U-net-Inc-AG-MS-DSLoss=FTL, gamma=1, Attention Gate=true, Multiscale Input=true, Deep Supervision=true, Incremental=true, Blob size=202021.06 | 64.5 | 90.9 | 9.14 | |
| Siamese U-netLoss=FTL, gamma=0.75, L2 regularization=10^-4, Blob size=202021.06 | 63.4 | 87.4 | 12.6 | |
| U-netLoss=FTL, gamma=1, Blob size=202021.06 | 61.8 | 96 | 4.04 | |
| U-netLoss=BCE, Blob size=202021.06 | 33 | 93.2 | 6.84 | |
| Li et al. 20182021.06 | 13.3 | 99.7 | 0.31 |