Research map / 3D vision / Depth and stereo
StruMonoNet: Structure-Aware Monocular 3D Prediction
Zhenpei Yang, Li Erran Li, Qixing Huang
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 |
|---|---|---|---|
| use StruMonoNet instead of Lee et al. | monocular depth prediction · NYUv2 | AbsRel: 0.110 -> 0.107 | Improved |
| use StruMonoNet instead of VPLNet | normal prediction · NYUv2 | mean angular error: 18.0 -> 16.8 | Improved |
| use StruMonoNet instead of Lee et al. | monocular depth prediction · ScanNet | AbsRel: 0.119 -> 0.103 | Improved |
| use StruMonoNet instead of Yu et al. | plane detection · ScanNet | Pixel Recall at 0.5: 0.550 -> 0.595 | Improved |
| add the structure prediction module (ablation) | StruMonoNet depth prediction · ScanNet | AbsRel error: 4.85 | Improved |
| add the structure prediction module (ablation) | StruMonoNet depth prediction · NYUv2 | AbsRel error: 5.62 | Improved |
| add pairwise plane-relation detection and enforcement (ablation) | StruMonoNet depth prediction · ScanNet | AbsRel error: 3.32 | Improved |
| add pairwise plane-relation detection and enforcement (ablation) | StruMonoNet depth prediction · NYUv2 | AbsRel error: 3.82 | Improved |
| enforce depth continuity between adjacent planes (ablation) | StruMonoNet depth prediction · ScanNet | AbsRel error: 2.02 | Improved |
| enforce depth continuity between adjacent planes (ablation) | StruMonoNet depth prediction · NYUv2 | AbsRel error: 2.12 | Improved |
| enforce geometric relations between planar normals (ablation) | StruMonoNet depth prediction · ScanNet | AbsRel error: 1.32 | Improved |
| enforce geometric relations between planar normals (ablation) | StruMonoNet depth prediction · NYUv2 | AbsRel error: 1.27 | Improved |
Closest papers by meaning
- MonoVLM: Monocular 3D Visual Grounding with Vision Language Models (CVPR 2026 · similarity 0.83)
- Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling (ECCV 2020 · similarity 0.83)
- Unsupervised Learning of 3D Structure from Images (NeurIPS 2016 · similarity 0.83)
- Last-Layer-Centric Feature Recombination: Unleashing 3D Geometric Knowledge in DINOv3 for Monocular Depth Estimation (ECCV 2026 · similarity 0.83)
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 2024 · similarity 0.83)
- S2R-DepthNet: Learning a Generalizable Depth-Specific Structural Representation (CVPR 2021 · similarity 0.83)
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.
