Research map / 3D vision / Face and shape
NeuS2: Fast Learning of Neural Implicit Surfaces for Multi-view Reconstruction
Yiming Wang, Qin Han, Marc Habermann, Kostas Daniilidis, Christian Theobalt, Lingjie Liu
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 progressive coarse-to-fine hash encoding training (ablation) | NeuS2 · DTU | Chamfer distance: 0.75 -> 0.70 | Improved |
| add Eikonal loss (ablation) | NeuS2 · DTU | Chamfer distance: 1.48 -> 0.70 | Improved |
| add global transformation prediction and progressive training (ablation) | NeuS2 dynamic · Dynamic ablation | LPIPS: 0.0595 -> 0.0474 | Improved |
| add global transformation prediction on top of progressive training (ablation) | NeuS2 dynamic · Dynamic ablation | PSNR: 27.72 -> 28.18 | Improved |
| add progressive training only (remove GTP) (ablation) | NeuS2 dynamic · Dynamic ablation | PSNR: 27.78 -> 27.72 | No clear effect |
| replace finite-difference second-order derivatives with analytic ReLU-MLP formulation (ablation) | NeuS2 · PyTorch comparison | PSNR: 33.4 -> 35.5 | Improved |
| use tailored second-order derivative formula instead of PyTorch autograd (ablation) | NeuS2 · PyTorch comparison | runtime: 10 min -> 7 min | Improved |
| use hash-encoded SDF with Eikonal loss instead of NeuS MLP | NeuS · DTU | runtime: 8 h -> 5 min | Improved |
| use NeuS2 instead of Instant-NGP for geometry (surface constraints) | Instant-NGP · DTU | Chamfer distance: 1.84 -> 0.70 | Improved |
| use NeuS2 instead of Instant-NGP for view synthesis | Instant-NGP · DTU | PSNR: 28.86 -> 28.82 | No clear effect |
| use NeuS2 incremental training instead of D-NeRF on real dynamic scenes | D-NeRF · Real dynamic scenes D1 | PSNR: 21.48 -> 26.41 | Improved |
Later papers that compared with this method
14 later accepted papers report NeuS2 as a baseline in their results tables; 11 of them report a better number on most of the results they share with it.
- Parametric SDF for Dynamic Surface Reconstruction (ECCV 2026 · better on 6 of 7 shared results)
- HoloTetSphere: Unified TetSphere Mesh Reconstruction for Physical Simulations (ECCV 2026 · better on 3 of 5 shared results)
- Hermite Radial Basis Function for Surface Reconstruction via Differentiable Rendering (CVPR 2026 · better on 2 of 4 shared results)
- Multi-view Surface Reconstruction Using Normal and Reflectance Cues (IJCV 2026 · better on 1 of 2 shared results)
- Mesh Splatting for End-to-end Multiview Surface Reconstruction (ICLR 2026 · better on 1 of 1 shared result)
- Ins-HOI: Instance Aware Human-Object Interactions Recovery (TPAMI 2025 · better on 6 of 6 shared results)
- MGSR: 2D/3D Mutual-boosted Gaussian Splatting for High-fidelity Surface Reconstruction under Various Light Conditions (ICCV 2025 · better on 0 of 2 shared results)
- ProbeSDF: Light Field Probes For Neural Surface Reconstruction (CVPR 2025 · better on 1 of 1 shared result)
6 more not listed.
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- D-NeRF — D-NeRF: Neural Radiance Fields for Dynamic Scenes (CVPR 2021) · better on 3 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
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
- SC-NeuS: Consistent Neural Surface Reconstruction from Sparse and Noisy Views (AAAI 2024 · similarity 0.90)
- NeuSurf: On-Surface Priors for Neural Surface Reconstruction from Sparse Input Views (AAAI 2024 · similarity 0.88)
- TFS-NeRF: Template-Free NeRF for Semantic 3D Reconstruction of Dynamic Scene (NeurIPS 2024 · similarity 0.87)
- UniSDF: Unifying Neural Representations for High-Fidelity 3D Reconstruction of Complex Scenes with Reflections (NeurIPS 2024 · similarity 0.87)
- ViiNeuS: Volumetric Initialization for Implicit Neural Surface Reconstruction of Urban Scenes with Limited Image Overlap (CVPR 2025 · similarity 0.86)
- From Transparent to Opaque: Rethinking Neural Implicit Surfaces with $\alpha$-NeuS (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.
