Segmentation on OASIS3 (test)
90.7Mdn DiceOurs w/o Lcs
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Ours w/o Lcstraining_set_proportion=100%, loss_config=without consistency loss2022.06 | 90.7 | — | — | 1.246 | 83.1 | 2.919 | 93.3 | — | — | |
| LNEtraining_set_proportion=100%2022.06 | 90.4 | — | — | 1.28 | 82.5 | 4.14 | 92.6 | — | — | |
| Ours w/ 0.1Lcstraining_set_proportion=100%, loss_config=consistency loss weight 0.12022.06 | 90.4 | — | — | 1.233 | 82.5 | 2.647 | 93.4 | — | — | |
| RandInitUnet.3Dtraining_set_proportion=100%, architecture=3D Unet2022.06 | 89.8 | — | — | 1.371 | 81.7 | 5.57 | 92 | — | — | |
| Ours w/ 1Lcstraining_set_proportion=100%, loss_config=consistency loss weight 1.02022.06 | 89.8 | — | — | 1.245 | 81.7 | 2.368 | 93.9 | — | — | |
| LNEtraining_mode=few-shot, finetuning_percentage=10%2022.06 | 89.4 | — | — | 1.426 | 81.1 | 2.747 | 92.3 | — | — | |
| Ours w/o Lcstraining_mode=few-shot, finetuning_percentage=10%, Lcs_weight=02022.06 | 89.4 | — | — | 1.372 | 81 | 2.647 | 92.3 | — | — | |
| Ours w/ 0.1Lcstraining_mode=few-shot, finetuning_percentage=10%, Lcs_weight=0.12022.06 | 89.4 | — | — | 1.335 | 80.9 | 2.602 | 92.6 | — | — | |
| PCLtraining_set_proportion=100%2022.06 | 89 | — | — | 1.315 | 80.2 | 3.328 | 92.6 | — | — | |
| Ours w/ 1Lcstraining_mode=few-shot, finetuning_percentage=10%, Lcs_weight=12022.06 | 89 | — | — | 1.341 | 80.4 | 2.446 | 93.2 | — | — | |
| PatchNCEtraining_set_proportion=100%2022.06 | 88.9 | — | — | 1.328 | 80.1 | 2.599 | 93.7 | — | — | |
| RandInitUnet.3Dtraining_mode=few-shot, finetuning_percentage=10%2022.06 | 88.8 | — | — | 1.409 | 80 | 2.575 | 92.6 | — | — | |
| PatchNCEtraining_mode=few-shot, finetuning_percentage=10%2022.06 | 88.8 | — | — | 1.347 | 80 | 2.666 | 92.7 | — | — | |
| GLCLtraining_set_proportion=100%2022.06 | 88.7 | — | — | 1.424 | 79.9 | 3.59 | 92.6 | — | — | |
| Context Restoretraining_set_proportion=100%2022.06 | 88.6 | — | — | 1.415 | 79.7 | 4.202 | 92.7 | — | — | |
| PCLtraining_mode=few-shot, finetuning_percentage=10%2022.06 | 88.5 | — | — | 1.328 | 79.6 | 3.896 | 91 | — | — | |
| GLCLtraining_mode=few-shot, finetuning_percentage=10%2022.06 | 88.4 | — | — | 1.387 | 79.4 | 4.005 | 91.1 | — | — | |
| RandInitUnet.2Dtraining_mode=few-shot, finetuning_percentage=10%2022.06 | 87.8 | — | — | 1.569 | 78.5 | 4.297 | 90.8 | — | — | |
| RandInitUnet.2Dtraining_set_proportion=100%, architecture=2D Unet2022.06 | 87.4 | — | — | 1.632 | 77.7 | 4.438 | 91.4 | — | — | |
| Context Restoretraining_mode=few-shot, finetuning_percentage=10%2022.06 | 87.3 | — | — | 1.66 | 77.7 | 4.528 | 91.6 | — | — | |
| AdvChainapproach=adversarial-based baseline2024.11 | — | 84.8 | — | — | — | — | — | — | — | |
| Finetuned SynthSegfinetuning=small number of real images2024.11 | — | 84.7 | — | — | — | — | — | — | — | |
| Learn2Synthsynthesis setting=parametric, noise parameter sigma=fixed2024.11 | — | 86 | — | — | — | — | — | — | — | |
| Learn2Synthsynthesis setting=parametric, noise parameter sigma=varying2024.11 | — | 85.7 | — | — | — | — | — | — | — | |
| Learn2Synthsynthesis setting=nonparametric, noise parameter sigma=fixed2024.11 | — | 87.5 | — | — | — | — | — | — | — | |
| Learn2Synthsynthesis setting=nonparametric, noise parameter sigma=varying2024.11 | — | 88.1 | — | — | — | — | — | — | — | |
| Mixed SynthSegaugmentation=small number of real images2024.11 | — | 85.4 | — | — | — | — | — | — | — | |
| Naive SynthSegtraining=from scratch in 2D2024.11 | — | 83.1 | — | — | — | — | — | — | — | |
| SAMSEGtype=unsupervised Bayesian2024.11 | — | 84.1 | — | — | — | — | — | — | — | |
| Supervised UNettraining_domain=real data2024.11 | — | 89.9 | — | — | — | — | — | — | — |