Research map / 3D vision / Face and shape
Sur^2f: A Hybrid Representation for High-Quality and Efficient Surface Reconstruction from Multi-view Images
Zhangjin Huang, Zhihao Liang, Kui Jia
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 |
|---|---|---|---|
| use surface-guided volume sampling instead of uniform sampling (ablation) | Sur^2f · DTU | PSNR: 30.09 -> 33.95 | Improved |
| use surface-guided sampling instead of hierarchical sampling (ablation) | Sur^2f · DTU | training time (s): 9538 -> 348 | Improved |
| use surface-guided sampling instead of uniform sampling (ablation) | Sur^2f · DTU | Chamfer Distance: 0.732 -> 0.661 | Improved |
| re-boot surrogate mesh with marching cubes every 500 iterations instead of every iteration (ablation) | Sur^2f · DTU | training time (s): 21869 -> 348 | Improved |
| increase re-boot interval from 500 to 1000 iterations (ablation) | Sur^2f · DTU | Chamfer Distance: 0.661 -> 0.715 | Hurt |
| use DMTet or FlexiCubes instead of marching cubes for re-booting (ablation) | Sur^2f · DTU | Chamfer Distance: 0.661 -> 0.662 | No clear effect |
| learn a shared neural shader for volume and surface rendering (ablation) | Sur^2f · DTU renderings | render consistency | Improved |
| use Sur^2f instead of DMTet/FlexiCubes in NVDIFFREC inverse rendering | NeRF synthetic Lego · NeRF synthetic Lego | PSNR: 29.0 -> 30.3 | Improved |
| use Sur^2f instead of DMTet/FlexiCubes in NVDIFFREC inverse rendering | NeRF synthetic Mats · NeRF synthetic Mats | PSNR: 27.0 -> 27.0 | No clear effect |
| train Sur^2f without foreground masks (ablation) | Sur^2f · DTU | Chamfer Distance: 0.66 -> 0.64 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- NeuS2 — NeuS2: Fast Learning of Neural Implicit Surfaces for Multi-view Reconstruction (ICCV 2023) · better on 2 of 3 shared results
- NeuS — NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction (NeurIPS 2021) · better on 3 of 3 shared results
- UNISURF — UNISURF: Unifying Neural Implicit Surfaces and Radiance Fields for Multi-View Reconstruction (ICCV 2021) · better on 2 of 2 shared results
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
- Learning Signed Distance Field for Multi-View Surface Reconstruction (ICCV 2021 · similarity 0.89)
- Critical Regularizations for Neural Surface Reconstruction in the Wild (CVPR 2022 · similarity 0.89)
- HybridNeRF: Efficient Neural Rendering via Adaptive Volumetric Surfaces (CVPR 2024 · similarity 0.88)
- ProbeSDF: Light Field Probes For Neural Surface Reconstruction (CVPR 2025 · similarity 0.88)
- 2S-UDF: A Novel Two-stage UDF Learning Method for Robust Non-watertight Model Reconstruction from Multi-view Images (CVPR 2024 · similarity 0.88)
- Neural Signed Distance Function Inference through Splatting 3D Gaussians Pulled on Zero-Level Set (NeurIPS 2024 · similarity 0.87)
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.
