Patient-level Burden Regression on AIBL (n=370)
0.42MAEMixMicrobleedNet
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MixMicrobleedNetGroup=Contextual reference2026.07 | 0.42 | 1.56 | 0.05 | |
| SHIVA-CMB v2Group=Contextual reference2026.07 | 0.87 | 2.75 | 0.51 | |
| DynUNetGroup=Baseline2026.07 | 0.93 | 1.84 | 0.53 | |
| CenSynCMBGroup=Our Method2026.07 | 1.87 | 4.12 | 0.43 | |
| Kim-style detectorGroup=Current methods2026.07 | 2.26 | 3.13 | 0.46 | |
| FRST-style detectorGroup=Current methods2026.07 | 4.68 | 6.11 | 0.47 | |
| Al-Masni-style detectorGroup=Current methods2026.07 | 6.09 | 7.16 | 0.41 | |
| Attention U-NetGroup=Baseline2026.07 | 6.68 | 8.1 | 0.41 | |
| Zihao VALDOGroup=VALDO methods2026.07 | 6.82 | 8.26 | 0.41 | |
| VISTA3DGroup=Baseline2026.07 | 7.43 | 9.19 | 0.41 | |
| SwinUNETRGroup=Baseline2026.07 | 10.26 | 11.82 | 0.34 |