Research map / 3D vision / Depth and stereo
Any to Full: Prompting Depth Anything for Depth Completion in One Stage
Zhiyuan Zhou, Ruofeng Liu, TAICHI LIU, Weijian Zuo, Shanshan Wang, Zhiqing Hong, Desheng Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add scale prompt (SP) fusion to Depth Anything-B (ablation) | Depth Anything-B · IBims-1 Hole | AbsREL: 0.047 -> 0.010 | Improved |
| add scale prompt (SP) fusion to Depth Anything-B (ablation) | Depth Anything-B · IBims-1 Range | AbsREL: 0.048 -> 0.068 | Hurt |
| add Local Enrichment module (ablation) | Any2Full (SP) · KITTI DC Range | AbsREL: 0.089 -> 0.092 | Hurt |
| add Local Enrichment module (ablation) | Any2Full (SP) · IBims-1 Hole | AbsREL: 0.010 -> 0.009 | Improved |
| add Global Propagation module (ablation) | Any2Full (SP+LE) · IBims-1 Range | AbsREL: 0.066 -> 0.031 | Improved |
| add Global Propagation module (ablation) | Any2Full (SP+LE) · KITTI DC Range | AbsREL: 0.092 -> 0.046 | Improved |
| add Global Propagation module (ablation) | Any2Full (SP+LE) · KITTI DC 4-line LiDAR | AbsREL: 0.065 -> 0.034 | Improved |
| use MDE-guided propagation instead of standard cross-attention (ablation) | Any2Full · IBims-1 Range | AbsREL: 0.057 -> 0.030 | Improved |
| add anchor loss to ssi, gm and r-ssim losses (ablation) | Any2Full DA-B · IBims-1 Range | AbsREL: 0.0348 -> 0.0306 | Improved |
| add r-ssim loss (ablation) | Any2Full DA-B · KITTI DC Range | AbsREL: 0.0600 -> 0.0632 | Hurt |
| use larger DA-L backbone instead of DA-B | Any2Full · six benchmarks avg | AbsREL: 0.010 vs 0.010 (Hole), 0.046 vs 0.046 (Range) | No clear effect |
| use MoGe-2-B backbone with scale prompting | MoGe-2-B · avg Hole | AbsREL: 0.059 -> 0.012 | Improved |
Closest papers by meaning
- Prompting Depth Anything for 4K Resolution Accurate Metric Depth Estimation (CVPR 2025 · similarity 0.90)
- A Simple yet Universal Framework for Depth Completion (NeurIPS 2024 · similarity 0.88)
- Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data (CVPR 2024 · similarity 0.88)
- Depth Anything V2 (NeurIPS 2024 · similarity 0.88)
- Depth Prompting for Sensor-Agnostic Depth Estimation (CVPR 2024 · similarity 0.87)
- Zero-shot Depth Completion via Test-time Alignment with Affine-invariant Depth Prior (AAAI 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
