3D on ETH3D
Metric: absolute relative error (lower is better)
Metric: absolute relative error (lower is better)
| # | Model↕ | absolute relative error▲ | Augmentations | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | HDN | 0.0121 | No | Hierarchical Normalization for Robust Monocular ... | 2022-10-18 | - |
| 2 | Depth Anything | 0.0127 | No | Depth Anything: Unleashing the Power of Large-Sc... | 2024-01-19 | Code |
| 3 | MiDaS 3.1 | 0.0139 | No | MiDaS v3.1 -- A Model Zoo for Robust Monocular R... | 2023-07-26 | Code |
| 4 | LeReS | 0.0171 | No | Learning to Recover 3D Scene Shape from a Single... | 2020-12-17 | Code |
| 5 | MiDaS | 0.0184 | No | Towards Robust Monocular Depth Estimation: Mixin... | 2019-07-02 | Code |
| 6 | DepthMaster | 0.053 | No | DepthMaster: Taming Diffusion Models for Monocul... | 2025-01-05 | Code |
| 7 | Distill Any Depth | 0.054 | No | Distill Any Depth: Distillation Creates a Strong... | 2025-02-26 | Code |
| 8 | Marigold | 0.065 | No | Repurposing Diffusion-Based Image Generators for... | 2023-12-04 | Code |
| 9 | PrimeDepth | 0.068 | No | PrimeDepth: Efficient Monocular Depth Estimation... | 2024-09-13 | Code |
| 10 | DPT | 0.078 | No | Vision Transformers for Dense Prediction | 2021-03-24 | Code |