Research map / 3D vision / Face and shape
UNISURF: Unifying Neural Implicit Surfaces and Radiance Fields for Multi-View Reconstruction
Michael Oechsle, Songyou Peng, Andreas Geiger
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 surface rendering and decaying sampling interval around the root-found surface (UNISURF) versus NeRF volume density | UNISURF · DTU (15 scans, mean) | Chamfer distance: 1.49 -> 1.02 | Improved |
| add unified occupancy formulation without masks versus mask-supervised IDR | UNISURF · DTU (15 scans, mean) | Chamfer distance: 0.90 -> 1.02 | No clear effect |
| add unified occupancy formulation versus COLMAP with trim 0 (watertight) | UNISURF · DTU (15 scans, mean) | Chamfer distance: 1.36 -> 1.02 | Improved |
| add volume rendering with decaying sampling interval to surface-rendering-only training (ablation) | UNISURF · DTU Buddha scene | reconstruction quality | Improved |
| add surface-guided sampling schedule versus uniform volume rendering with 96 points (ablation) | UNISURF · DTU Buddha scene | reconstruction quality | Improved |
| add surface-guided sampling versus NeRF hierarchical sampling (64+64) (ablation) | UNISURF · DTU Buddha scene | reconstruction quality | Improved |
| add normal-smoothness surface regularizer (ablation) | UNISURF · DTU Buddha scene | surface smoothness | Improved |
| render with surface rendering instead of volume rendering at inference | UNISURF · novel view synthesis | inference speed: twice as fast | Improved |
Later papers that compared with this method
19 later accepted papers report Unisurf as a baseline in their results tables; 17 of them report a better number on most of the results they share with it.
- VRP-UDF: Toward Unbiased Learning of Unsigned Distance Functions From Multi-View Images With Volume Rendering Priors (TPAMI 2026 · better on 1 of 1 shared result)
- Free-Moving Object Reconstruction and Pose Estimation with Virtual Camera (AAAI 2025 · better on 1 of 1 shared result)
- Sur^2f: A Hybrid Representation for High-Quality and Efficient Surface Reconstruction from Multi-view Images (ECCV 2024 · better on 2 of 2 shared results)
- DebSDF: Delving Into the Details and Bias of Neural Indoor Scene Reconstruction (TPAMI 2024 · better on 2 of 2 shared results)
- In-Hand 3D Object Scanning From an RGB Sequence (CVPR 2023 · better on 2 of 4 shared results)
- Learning Shape Primitives via Implicit Convexity Regularization (ICCV 2023 · better on 1 of 4 shared results)
- CorresNeRF: Image Correspondence Priors for Neural Radiance Fields (NeurIPS 2023 · better on 2 of 2 shared results)
- Looking Through the Glass: Neural Surface Reconstruction Against High Specular Reflections (CVPR 2023 · better on 1 of 1 shared result)
11 more not listed.
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- UNIS: A Unified Framework for Achieving Unbiased Neural Implicit Surfaces in Volume Rendering (ICCV 2025 · similarity 0.92)
- Neural RGB-D Surface Reconstruction (CVPR 2022 · similarity 0.91)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (CVPR 2025 · similarity 0.91)
- Towards Unbiased Volume Rendering of Neural Implicit Surfaces With Geometry Priors (CVPR 2023 · similarity 0.91)
- ProbeSDF: Light Field Probes For Neural Surface Reconstruction (CVPR 2025 · similarity 0.89)
- Differentiable Neural Surface Refinement for Modeling Transparent Objects (CVPR 2024 · similarity 0.89)
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.
