Research map / 3D vision / Depth and stereo
SURGE: Surface Regularized Geometry Estimation from a Single Image
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 DCRF regularization with planar affinity and depth-normal compatibility (with confidence and joint training) | Eigen-VGG normals · NYU v2 planar regions | plane-wise degree variation (var.): 9.1534 -> 6.3939 | Helped |
| Add DCRF regularization (with confidence and joint training) | Eigen-VGG normals · NYU v2 planar regions | mean normal error: 14.5425 -> 13.9763 | Helped |
| Hard planar post-processing of normals (ablation) | Eigen-VGG normals · NYU v2 planar regions | mean normal error: 14.5425 -> 14.8154 | Hurt |
| Hard planar post-processing of normals (ablation) | Eigen-VGG normals · NYU v2 planar regions | plane-wise degree variation (var.): 9.1534 -> 7.2753 | Helped |
| Joint training of CNN with DCRF (CNN only) (ablation) | Eigen-VGG normals · NYU v2 planar regions | mean normal error: 14.5425 -> 14.4978 | Helped |
| Add DCRF regularization without confidence weighting (ablation) | Eigen-VGG normals · NYU v2 planar regions | plane-wise degree variation (var.): 9.1534 -> 6.9688 | Helped |
| Joint training of DCRF and CNN on top of DCRF (depth) | Eigen-VGG depth · NYU v2 planar regions | plane-wise degree variation (var.) depth: 16.4460 -> 10.5746 | Helped |
| Hard planar post-processing of depth (ablation) | Eigen-VGG depth · NYU v2 planar regions | Rel: 0.1441 -> 0.1470 | Hurt |
| Add DCRF regularization with joint training (depth) | Eigen-VGG depth · NYU v2 planar regions | Rel: 0.1441 -> 0.1423 | Helped |
| Add DCRF regularization on normals (edge region) | Eigen-VGG normals · NYU v2 edge region | mean normal error: 23.4141 -> 23.4694 | Hurt |
| Add DCRF regularization on normals (edge region) | Eigen-VGG normals · NYU v2 edge region | median normal error: 18.3288 -> 17.6804 | Helped |
| Add DCRF regularization on normals (full image) | Eigen-VGG normals · NYU v2 full image | mean normal error: 20.9322 -> 20.6093 | Helped |
Closest papers by meaning
- SurGe: Improved Surface Geometry in Point Maps (NeurIPS 2026 · similarity 0.88)
- Marr Revisited: 2D-3D Alignment via Surface Normal Prediction (CVPR 2016 · similarity 0.88)
- Deep Optimized Priors for 3D Shape Modeling and Reconstruction (CVPR 2021 · similarity 0.87)
- Unsupervised Learning of Geometry From Videos With Edge-Aware Depth-Normal Consistency (AAAI 2018 · similarity 0.87)
- VCR-GauS: View Consistent Depth-Normal Regularizer for Gaussian Surface Reconstruction (NeurIPS 2024 · similarity 0.86)
- Flatten Anything: Unsupervised Neural Surface Parameterization (NeurIPS 2024 · 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.