HAM10000
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Skin lesion classification | HAM10000 (test) | Accuracy93.32 | 83 | |
| Medical Image Classification | HAM10000 | Accuracy83.4 | 39 | |
| Lesion Segmentation | HAM10000 | HD956.45 | 38 | |
| Skin Lesion Segmentation | HAM10000 | Dice Coefficient95.16 | 34 | |
| Classification | HAM10000 | F1 Score95.83 | 30 | |
| Medical Image Segmentation | HAM10000 | mDSC0.9471 | 27 | |
| Image Classification | HAM10000 (test) | Accuracy65.63 | 24 | |
| Classification | HAM10000+ | Accuracy84.48 | 22 | |
| Segmentation | HAM10000 (Hard Samples) | IoU79.92 | 21 | |
| Skin lesion classification | HAM10000 | Accuracy88 | 20 | |
| Image Denoising | HAM10000 (test) | SSIM0.95 | 20 | |
| Image Denoising | HAM10000 52 (test) | PSNR38.8 | 20 | |
| Medical Image Classification | HAM10000 (random) | Accuracy83.3 | 19 | |
| Image Classification | HAM10000 | Accuracy97.1 | 19 | |
| Lesion Segmentation | HAM10000 Domain Generalization train on NV unseen categories (test) | AKIEC IoU74.53 | 16 | |
| Medical Grounding | HAM10000 | A@0.597.2 | 16 | |
| Attribution Analysis | HAM10000 | SEG-GRAD0.1726 | 13 | |
| Image Classification | HAM10000 NIID, κ=1 | Accuracy70.46 | 12 | |
| Image Classification | HAM10000 Uniform | Accuracy78.51 | 12 | |
| Categorical skin lesion classification | HAM10000 original (test) | Accuracy0.8915 | 12 | |
| Medical Image Classification | HAM10000 D5 (test) | Accuracy99.8 | 12 | |
| Imbalanced Classification | HAM10000 (test) | Macro F1 Score81.97 | 10 | |
| Skin Lesion Classification | HAM10000 NID2 label quantity skew | Accuracy46.17 | 9 | |
| Skin Lesion Classification | HAM10000 NID1_0.5 label/quantity skew | Accuracy52.79 | 9 | |
| Skin Lesion Classification | HAM10000 NID1_0.2 label/quantity skew | Accuracy50.23 | 9 |
Showing 25 of 58 rows