Image-to-text retrieval on Open-I
0.449mAPDDBH
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DDBHHash Bits=64, Backbone=BioMedCLIP2025.12 | 0.449 | — | — | — | |
| MCMFHHash Bits=32, Backbone=BioMedCLIP2025.12 | 0.436 | — | — | — | |
| MCMFHHash Bits=64, Backbone=BioMedCLIP2025.12 | 0.434 | — | — | — | |
| DDBHHash Bits=32, Backbone=BioMedCLIP2025.12 | 0.432 | — | — | — | |
| MCMFHHash Bits=16, Backbone=BioMedCLIP2025.12 | 0.429 | — | — | — | |
| MCMFH w/SwitchHash Bits=162025.12 | 0.421 | — | — | — | |
| MCMFH w/Variance-basedHash Bits=162025.12 | 0.419 | — | — | — | |
| MCMFH +MoEHash Bits=162025.12 | 0.412 | — | — | — | |
| MCMFH +MoE+Voting5+UnfrozenHash Bits=162025.12 | 0.404 | — | — | — | |
| DDBHHash Bits=16, Backbone=BioMedCLIP2025.12 | 0.401 | — | — | — | |
| MCMFH +MoE+Voting1+FrozenHash Bits=162025.12 | 0.383 | — | — | — | |
| UCMFHHash Bits=64, Backbone=BioMedCLIP2025.12 | 0.382 | — | — | — | |
| UCMFHHash Bits=32, Backbone=BioMedCLIP2025.12 | 0.367 | — | — | — | |
| UCMFHHash Bits=16, Backbone=BioMedCLIP2025.12 | 0.354 | — | — | — | |
| DScPHHash Bits=64, Backbone=BioMedCLIP2025.12 | 0.28 | — | — | — | |
| DScPHHash Bits=32, Backbone=BioMedCLIP2025.12 | 0.247 | — | — | — | |
| DScPHHash Bits=16, Backbone=BioMedCLIP2025.12 | 0.194 | — | — | — | |
| CLIP2024.03 | — | 0.8 | — | 8.1 | |
| CXR-CLIPBackbone=ResNet-50, Pre-train Dataset=MIMIC-CXR2023.10 | — | 3.8 | 8.2 | 12.3 | |
| CXR-CLIPBackbone=Swin-Tiny, Pre-train Dataset=MIMIC-CXR2023.10 | — | 3.6 | 8.3 | 11.5 | |
| CXR-CLIPBackbone=ResNet-50, Pre-train Dataset=MIMIC-CXR, CheXpert2023.10 | — | 3.5 | 8.2 | 12 | |
| CXR-CLIPBackbone=Swin-Tiny, Pre-train Dataset=MIMIC-CXR, CheXpert2023.10 | — | 3.1 | 8.3 | 11.6 | |
| CXR-CLIPBackbone=ResNet-50, Pre-train Dataset=MIMIC-CXR, CheXpert, Chest X-ray142023.10 | — | 2.3 | 6.7 | 10.1 | |
| CXR-CLIPBackbone=Swin-Tiny, Pre-train Dataset=MIMIC-CXR, CheXpert, Chest X-ray142023.10 | — | 2.4 | 6.6 | 9.4 | |
| CXR-CLIP2024.03 | — | 14.1 | — | 39.3 | |
| GloRIABackbone=ResNet-50, Pre-train Dataset=CheXpert with reports2023.10 | — | 1.5 | 4.4 | 6.5 | |
| MEDBindvariant=NM2024.03 | — | 14.3 | — | 41.1 | |
| MEDBindvariant=BD2024.03 | — | 13.6 | — | 39.7 | |
| MedCLIPBackbone=Swin-Tiny, Pre-train Dataset=MIMIC-CXR, CheXpert2023.10 | — | 0.1 | 0.4 | 0.7 | |
| MedCLIP2024.03 | — | 0.9 | — | 8.6 |