Linear Classification on COVIDx (test)
93AccuracySISTA
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SISTABackbone=ViT-B/16, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 93 | — | — | — | |
| MLIPBackbone=ViT-B/16, Venue=CVPR’24, incorporate auxiliary knowledge=true, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 92.5 | — | — | — | |
| MGCABackbone=ViT-B/16, Venue=NIPS’22, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 92.3 | — | — | — | |
| ConVIRTBackbone=RN-50, Venue=PMLR’22, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 92 | — | — | — | |
| SISTABackbone=RN-50, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 92 | — | — | — | |
| PRIORBackbone=RN-50, Venue=ICCV’23, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 91 | — | — | — | |
| M-FLAGBackbone=RN-50, Venue=MICCAI’23, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 90.7 | — | — | — | |
| MGCABackbone=RN-50, Venue=NIPS’22, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 90.5 | — | — | — | |
| MLIPBackbone=RN-50, Venue=CVPR’24, incorporate auxiliary knowledge=true, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 90.5 | — | — | — | |
| MedKLIPBackbone=RN-50, Venue=ICCV’23, incorporate instance-level labels=true, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 90.3 | — | — | — | |
| GLoRIABackbone=RN-50, Venue=ICCV’21, Training data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 88.6 | — | — | — | |
| SISTABackbone=ViT-B/16, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 86.5 | — | — | — | |
| ImageNetTraining data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 86.3 | — | — | — | |
| MLIPBackbone=ViT-B/16, Venue=CVPR’24, incorporate auxiliary knowledge=true, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 86.3 | — | — | — | |
| SISTABackbone=RN-50, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 86 | — | — | — | |
| MedKLIPBackbone=RN-50, Venue=ICCV’23, incorporate instance-level labels=true, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 85.2 | — | — | — | |
| MLIPBackbone=RN-50, Venue=CVPR’24, incorporate auxiliary knowledge=true, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 85 | — | — | — | |
| MGCABackbone=ViT-B/16, Venue=NIPS’22, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 84.8 | — | — | — | |
| PRIORBackbone=RN-50, Venue=ICCV’23, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 84.7 | — | — | — | |
| M-FLAGBackbone=RN-50, Venue=MICCAI’23, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 84.1 | — | — | — | |
| MGCABackbone=RN-50, Venue=NIPS’22, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 83.5 | — | — | — | |
| ConVIRTBackbone=RN-50, Venue=PMLR’22, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 82.5 | — | — | — | |
| GLoRIABackbone=RN-50, Venue=ICCV’21, Training data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 81.5 | — | — | — | |
| ImageNetTraining data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 78.8 | — | — | — | |
| MLIPBackbone=ViT-B/16, Venue=CVPR’24, incorporate auxiliary knowledge=true, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 75.3 | — | — | — | |
| SISTABackbone=ViT-B/16, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 75 | — | — | — | |
| MGCABackbone=ViT-B/16, Venue=NIPS’22, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 74.8 | — | — | — | |
| MedKLIPBackbone=RN-50, Venue=ICCV’23, incorporate instance-level labels=true, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 74.5 | — | — | — | |
| MLIPBackbone=RN-50, Venue=CVPR’24, incorporate auxiliary knowledge=true, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 73 | — | — | — | |
| ConVIRTBackbone=RN-50, Venue=PMLR’22, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 72.5 | — | — | — | |
| PRIORBackbone=RN-50, Venue=ICCV’23, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 72.3 | — | — | — | |
| SISTABackbone=RN-50, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 72.3 | — | — | — | |
| M-FLAGBackbone=RN-50, Venue=MICCAI’23, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 72.2 | — | — | — | |
| MGCABackbone=RN-50, Venue=NIPS’22, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 72 | — | — | — | |
| RandomTraining data percentage=100%, Evaluation protocol=Fine-tuning2026.01 | 70 | — | — | — | |
| GLoRIABackbone=RN-50, Venue=ICCV’21, Training data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 67.3 | — | — | — | |
| ImageNetTraining data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 64.8 | — | — | — | |
| RandomTraining data percentage=10%, Evaluation protocol=Fine-tuning2026.01 | 60.3 | — | — | — | |
| RandomTraining data percentage=1%, Evaluation protocol=Fine-tuning2026.01 | 50.5 | — | — | — | |
| ConVIRTPre-training=MIMIC-CXR2022.10 | — | 72.5 | 82.5 | 92 | |
| GLoRIAPre-training=CheXpert2022.10 | — | 67.3 | 77.8 | 89 | |
| GLoRIA-MIMICPre-training=MIMIC-CXR2022.10 | — | 66.5 | 80.5 | 88.8 | |
| ImageNet InitPre-training=ImageNet2022.10 | — | 64.8 | 78.8 | 86.3 | |
| MGCAPre-training=MIMIC-CXR, Backbone=ResNet-502022.10 | — | 72 | 83.5 | 90.5 | |
| MGCAPre-training=MIMIC-CXR, Backbone=ViT-B/162022.10 | — | 74.8 | 84.8 | 92.3 | |
| Random InitPre-training=None2022.10 | — | 50.5 | 60.3 | 70 |