Research map / 3D vision / Depth and stereo
Deep Surface Normal Estimation With Hierarchical RGB-D Fusion
Jin Zeng, Yanfeng Tong, Yunmu Huang, Qiong Yan, Wenxiu Sun, Jing Chen, Yongtian Wang
Check 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 |
|---|---|---|---|
| replace single-scale early or late fusion with hierarchical fusion (ablation) | HFM-Net · Matterport3D and ScanNet | mean angular error: 13.968 -> 13.062; 13.645 -> 13.062; 16.045 -> 14.590; 17.425 -> 14.590 | Improved |
| replace binary mask with a learned confidence map (ablation) | hierarchical fusion with hybrid loss · Matterport3D and ScanNet | mean angular error: 13.437 -> 13.062; 14.696 -> 14.590 | Improved |
| use hybrid multi-scale loss instead of L2 loss at all scales (ablation) | hierarchical fusion with confidence map · Matterport3D and ScanNet | mean angular error: 13.688 -> 13.062; 14.946 -> 14.590 | Improved |
| combine RGB and depth using HFM-Net instead of GFMM | surface normal estimation · Matterport3D | mean angular error: 16.537 -> 13.062 | Improved |
| combine RGB and depth using HFM-Net instead of GFMM | surface normal estimation · ScanNet | mean angular error: 21.174 -> 14.590 | Improved |
Closest papers by meaning
- Deep Surface Normal Estimation on the 2-Sphere with Confidence Guided Semantic Attention (ECCV 2020 · similarity 0.88)
- VPLNet: Deep Single View Normal Estimation With Vanishing Points and Lines (CVPR 2020 · similarity 0.87)
- Normal Assisted Stereo Depth Estimation (CVPR 2020 · similarity 0.87)
- Adaptive Surface Normal Constraint for Depth Estimation (ICCV 2021 · similarity 0.86)
- Just Look at the Image: Viewpoint-Specific Surface Normal Prediction for Improved Multi-View Reconstruction (CVPR 2016 · similarity 0.86)
- Hierarchical Normalization for Robust Monocular Depth Estimation (NeurIPS 2022 · 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.
