Downstream classification on BraTS 2019
71AccuracyMulti-Scale Reward Learning
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Multi-Scale Reward LearningClassifier backbone=3D ResNet-50, Evaluation protocol=Pre-train on synthetic, fine-tune on real, Number of independent runs=32026.03 | 71 | 71 | 74 | |
| TAMTClassifier backbone=3D ResNet-50, Evaluation protocol=Pre-train on synthetic, fine-tune on real, Number of independent runs=32026.03 | 67 | 71 | 77 | |
| 3D-Med-DDPMClassifier backbone=3D ResNet-50, Evaluation protocol=Pre-train on synthetic, fine-tune on real, Number of independent runs=32026.03 | 61 | 64 | 69 | |
| 3D-αWGANClassifier backbone=3D ResNet-50, Evaluation protocol=Pre-train on synthetic, fine-tune on real, Number of independent runs=32026.03 | 59 | 59 | 70 |