Monocular Depth Estimation on SCARED
0.051Abs RelDV-LORA
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DV-LORA2025.12 | 0.051 | 0.36 | 4.527 | 0.073 | 98.1 | — | |
| EndoDAC2025.12 | 0.052 | 0.37 | 4.582 | 0.074 | 97.6 | — | |
| DAEvaluation Protocol=fine-tuned2025.12 | 0.058 | 0.451 | 5.058 | 0.081 | 97.4 | — | |
| AF-SfMLearner2025.12 | 0.059 | 0.435 | 4.925 | 0.082 | 97.4 | — | |
| Endo-SfM2025.12 | 0.062 | 0.606 | 5.726 | 0.093 | 95.7 | — | |
| [32]2025.12 | 0.062 | 0.558 | 5.585 | 0.09 | 96.2 | — | |
| SC-SfMLearner2025.12 | 0.068 | 0.645 | 5.988 | 0.097 | 95.7 | — | |
| Monodepth22025.12 | 0.069 | 0.577 | 5.546 | 0.094 | 94.8 | — | |
| Defeat-Net2025.12 | 0.077 | 0.792 | 6.688 | 0.108 | 94.1 | — | |
| [28]2025.12 | 0.078 | 0.794 | 6.794 | 0.109 | 94.6 | — | |
| DAEvaluation Protocol=zero-shot2025.12 | 0.084 | 0.847 | 6.711 | 0.11 | 93 | — | |
| ZENBackbone=ViT-B, Fine-tuned=true, Pre-training strategy=Existing pretrained models2026.02 | 0.1306 | 1.5402 | 7.9178 | 0.1386 | — | 85.92 | |
| MIS-MoCoV3Backbone=ViT-B, Fine-tuned=true, Pre-training strategy=Self-supervised learning2026.02 | 0.1326 | 1.6049 | 8.2549 | 0.144 | — | 83.25 | |
| SurgeNetBackbone=CAFormer18, Fine-tuned=true, Pre-training strategy=Existing pretrained models2026.02 | 0.1329 | 1.51 | 8.1794 | 0.1438 | — | 84.73 | |
| MIS-MSNBackbone=ViT-B, Fine-tuned=true, Pre-training strategy=Self-supervised learning2026.02 | 0.1333 | 1.6449 | 8.3485 | 0.1451 | — | 84.07 | |
| MIS-DINOBackbone=ViT-B, Fine-tuned=true, Pre-training strategy=Self-supervised learning2026.02 | 0.1361 | 1.6881 | 8.5661 | 0.1492 | — | 82.71 | |
| ImageNetBackbone=ViT-B, Fine-tuned=true, Pre-training strategy=Existing pretrained models2026.02 | 0.1364 | 1.7377 | 8.5183 | 0.148 | — | 82.99 | |
| EndoFMBackbone=ViT-B, Fine-tuned=true, Pre-training strategy=Existing pretrained models2026.02 | 0.1365 | 1.7387 | 8.5228 | 0.1491 | — | 82.78 | |
| MIS-DINOv2Backbone=ViT-B, Fine-tuned=true, Pre-training strategy=Self-supervised learning2026.02 | 0.1385 | 1.7175 | 8.4867 | 0.1491 | — | 81.93 | |
| MIS-MAEBackbone=ViT-B, Fine-tuned=true, Pre-training strategy=Self-supervised learning2026.02 | 0.1416 | 1.9033 | 8.8798 | 0.1544 | — | 81.63 | |
| MIS-DINOv2Backbone=ViT-L, Fine-tuned=true, Pre-training strategy=Self-supervised learning2026.02 | 0.1417 | 1.7527 | 8.762 | 0.1529 | — | 82.14 | |
| EndoViTBackbone=ViT-B, Fine-tuned=true, Pre-training strategy=Existing pretrained models2026.02 | 0.1477 | 2.1513 | 9.0645 | 0.1574 | — | 81.42 | |
| PeskaVLPBackbone=ResNet50, Fine-tuned=true, Pre-training strategy=Existing pretrained models2026.02 | 0.1545 | 2.126 | 9.9993 | 0.1748 | — | 78.35 | |
| GSViTBackbone=EfficientViT, Fine-tuned=true, Pre-training strategy=Existing pretrained models2026.02 | 0.2246 | 4.3211 | 13.3113 | 0.2337 | — | 62.81 | |
| Leveling3D2026.03 | 1.117 | — | 19.05 | 0.818 | 23.4 | — | |
| Difix3D+2026.03 | 1.133 | — | 19.19 | 0.824 | 23.2 | — | |
| Baseline2026.03 | 1.192 | — | 20.25 | 3.7 | 19.5 | — |