Affine-invariant Depth Estimation on DIODE Full
0.136AbsRelOurs CSTM_label (ViT-giant)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Ours CSTM_label (ViT-giant)Backbone=ViT-giant, #Params=1011M, Pretrain=142M, #Data Train=16M, Zero-shot transfer=true2024.03 | 0.136 | 0.895 | |
| Ours CSTM_label (ViT-Large)Backbone=ViT-Large, #Params=306M, Pretrain=142M, #Data Train=16M, Zero-shot transfer=true2024.03 | 0.141 | 0.882 | |
| DPT-largeBackbone=ViT-Large, Pretrain=1.3M, #Data Train=188K, Zero-shot transfer=true2024.03 | 0.182 | 0.78 | |
| Ours CSTM_image (ConvNeXt-Large)Backbone=ConvNeXt-Large, #Params=198M, Pretrain=14.2M, #Data Train=8M, Zero-shot transfer=true2024.03 | 0.211 | 0.825 | |
| Ours CSTM_label (ConvNeXt-Large)Backbone=ConvNeXt-Large, #Params=198M, Pretrain=14.2M, #Data Train=8M, Zero-shot transfer=true2024.03 | 0.224 | 0.805 | |
| HDNBackbone=ViT-Large, #Params=306M, Pretrain=1.3M, #Data Train=300K, Zero-shot transfer=true2024.03 | 0.246 | 0.882 | |
| Ours CSTM_label (ViT-Small)Backbone=ViT-Small, #Params=22M, Pretrain=142M, #Data Train=16M, Zero-shot transfer=true2024.03 | 0.247 | 0.789 | |
| LeresBackbone=ResNeXt101, Pretrain=1.3M, #Data Train=354K, Zero-shot transfer=true2024.03 | 0.271 | 0.766 | |
| DepthAnythingBackbone=ViT-Large, #Params=142M, Pretrain=142M, #Data Train=63.5M, Zero-shot transfer=true2024.03 | 0.277 | 0.759 | |
| MarigoldBackbone=Latent diffusion V2, #Params=899M, Pretrain=5B, #Data Train=74K, Zero-shot transfer=true2024.03 | 0.308 | 0.773 | |
| MiDaSBackbone=ResNeXt101, #Params=88M, Pretrain=1.3M, #Data Train=2M, Zero-shot transfer=true2024.03 | 0.332 | 0.715 | |
| OmnidataBackbone=ViT-Base, Pretrain=1.3M, #Data Train=12.2M, Zero-shot transfer=true2024.03 | 0.339 | 0.742 | |
| DiverseDepthBackbone=ResNeXt50, #Params=25M, Pretrain=1.3M, #Data Train=320K, Zero-shot transfer=true2024.03 | 0.376 | 0.631 | |
| DPTTraining Data=1.4M, Zero-shot=true2026.03 | 18.2 | 0.758 | |
| Lotus-D*Training Data=59K, Zero-shot=true2026.03 | 22.8 | 0.738 | |
| GenPercept*Training Data=90K, Zero-shot=true2026.03 | 22.8 | 0.74 | |
| Lotus-G*Training Data=59K, Zero-shot=true2026.03 | 22.9 | 0.729 | |
| IrisTraining Data=59K + 100K, Zero-shot=true2026.03 | 24.3 | 0.743 | |
| HDNTraining Data=300K, Zero-shot=true2026.03 | 24.6 | 0.78 | |
| DepthAnythingTraining Data=62.6M, Zero-shot=true2026.03 | 26 | 0.759 | |
| DepthAnything V2Training Data=62.6M, Zero-shot=true2026.03 | 26.5 | 0.734 | |
| LeResTraining Data=354K, Zero-shot=true2026.03 | 27.1 | 0.766 | |
| Diffusion-E2E-FT*Training Data=74K, Zero-shot=true2026.03 | 30.3 | 0.776 | |
| Marigold(LCM)*Training Data=74K, Zero-shot=true2026.03 | 30.7 | 0.775 | |
| Marigold*Training Data=74K, Zero-shot=true2026.03 | 30.8 | 0.773 | |
| MiDaSTraining Data=2M, Zero-shot=true2026.03 | 33.2 | 0.715 | |
| GeoWizard*Training Data=280K, Zero-shot=true2026.03 | 33.5 | 0.723 | |
| OmnidataTraining Data=12.2M, Zero-shot=true2026.03 | 33.9 | 0.742 | |
| DiverseDepthTraining Data=320K, Zero-shot=true2026.03 | 37.6 | 0.631 |