Multi-label Medical Image Classification on MIMIC Internal (test)
0.83AUROC (ATEL)MedSigLIP(n=33M)
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MedSigLIP(n=33M)LP Training Data=2D, Evaluation Protocol=Linear Probing2026.05 | 0.83 | 0.83 | 0.85 | 0.79 | 0.91 | 0.84 | |
| RAD-DINO(n=0.84M)LP Training Data=2D, Evaluation Protocol=Linear Probing2026.05 | 0.82 | 0.82 | 0.83 | 0.78 | 0.91 | 0.83 | |
| MultiMedVision-S3LP Training Data=ALL, Evaluation Protocol=Linear Probing2026.05 | 0.81 | 0.82 | 0.81 | 0.76 | 0.9 | 0.82 | |
| BiomedCLIP(n=15M)LP Training Data=2D, Evaluation Protocol=Linear Probing2026.05 | 0.78 | 0.8 | 0.82 | 0.75 | 0.88 | 0.81 | |
| SupervisedLP Training Data=ALL, Evaluation Protocol=Supervised2026.05 | 0.78 | 0.79 | 0.77 | 0.71 | 0.85 | 0.78 |