ODIR
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Image Classification | ODIR 200x3 | ACA89.3 | 60 | |
| Medical Classification | ODIR (3-fold cross-validation) | Accuracy (ACC)80.61 | 48 | |
| Disease diagnosis | ODIR 21 (test) | F1 Macro99.77 | 36 | |
| Medical Image Re-identification | ODIR Fundus | CMC-R171.34 | 29 | |
| Multi-disease Classification | ODIR5K Full set Few-shot | AUROC71.65 | 16 | |
| Medical image classification | ODIR-5k | Accuracy69.6 | 16 | |
| Multi-label fundus diagnosis | ODIR-5K | AUC-ROC87.23 | 15 | |
| Image-to-Text Retrieval | ODIR-5K (test) | mAP (64 bits)97.1 | 14 | |
| Classification | ODIR 200x3 CFP Modality | Accuracy86.3 | 8 | |
| Fundus Image Classification | ODIR to VietAI (target) | Avg Score80.8 | 8 | |
| Medical image classification | ODIR-5k (test) | ACC69.6 | 8 | |
| Ocular Disease Recognition | ODIR-5K | AUC-ROC81.5 | 7 | |
| Multimodal Diagnosis | ODIR | Parameters1.61 | 6 | |
| Ocular Disease Recognition | ODIR (test) | AUC0.865 | 6 | |
| Open-ended Visual Question Answering | ODIR | F1 Score56.7 | 6 | |
| Closed Visual Question Answering | ODIR | F1 Score60.09 | 6 | |
| Concept faithfulness | ODIR-5K retinal fundus mean over four disease classes | Accuracy100 | 4 |
Showing 17 of 17 rows