Multi-view depth estimation on Tanks and Temples (test)
2.04Relative ErrorMASt3R
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MASt3REncoder=ViT-Large2025.03 | 2.04 | 87.88 | |
| DUNEEncoder=ViT-Base2025.03 | 2.6 | 79.19 | |
| DUSt3REncoder=ViT-Large2025.03 | 2.93 | 78.51 | |
| DUSt3R-224Resolution=224, In-domain training=false, Inference Time (s)=0.12025.03 | 3.7 | 68.2 | |
| Spann3RResolution=224, In-domain training=false, Inference Time (s)=0.292025.03 | 3.7 | 66.2 | |
| DUSt3R-512Resolution=512, In-domain training=false, Inference Time (s)=0.192025.03 | 4.6 | 63.1 | |
| MUSt3R-224Resolution=224, In-domain training=false, Inference Time (s)=0.322025.03 | 5.3 | 57.4 | |
| MUSt3R-512Resolution=512, In-domain training=false, Inference Time (s)=0.192025.03 | 5.7 | 58.6 | |
| DeepV2DEncoder=Hourglass2025.03 | 8.9 | 46.4 |