Research map / 3D vision / Depth and stereo
ManifoldNeuS: Manifold-aware View Optimizability for Pose-Free Neural Surface Reconstruction
Xinxin Liu, Xue Wang, Guoqing Zhou, Qing Wang
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 MaVOS-based view scheduling to naive NeuS-BARF (ablation) | NeuS-BARF · DTU Scan24 | rotation error (deg): 168.05 -> 36.02 | Improved |
| add MaVOS-gated positional encoding (ablation) | ManifoldNeuS · DTU Scan24 | rotation error (deg): 1.80 -> 0.80 | Improved |
| add anti-score loss weighting instead of base weighting (ablation) | ManifoldNeuS · DTU Scan24 | rotation error (deg): 2.09 -> 0.80 | Improved |
| add manifold coverage term to MaVOS (ablation) | ManifoldNeuS · DTU Scan24 | rotation error (deg): 1.38 -> 0.80 | Improved |
| add dynamic refinement (reordering successors) to view scheduling (ablation) | ManifoldNeuS · DTU Scan24 | rotation error (deg): 4.07 -> 0.80 | Improved |
| use ManifoldNeuS instead of VGGT for pose estimation | pose-free reconstruction · DTU (9 scenes) average | rotation error (deg): 2.300 -> 0.602 | Improved |
| use ManifoldNeuS instead of VGGT for pose estimation on Scan24 | pose-free reconstruction · DTU Scan24 | rotation error (deg): 0.516 -> 0.795 | Hurt |
| use ManifoldNeuS identity-initialized instead of NeuS-BARF | pose-free NeuS · DTU (9 scenes) average | rotation error (deg): 118.475 -> 0.602 | Improved |
| use ManifoldNeuS pose-free instead of NeuS with COLMAP poses | surface reconstruction · DTU (9 scenes) average | Chamfer distance: 0.368 -> 0.404 | Hurt |
| use ManifoldNeuS instead of SG-NeRF for reconstruction | surface reconstruction · DTU (9 scenes) average | Chamfer distance: 0.590 -> 0.404 | Improved |
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.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 3 of 4 shared results
- SG-NeRF — SG-NeRF: Neural Surface Reconstruction with Scene Graph Optimization (ECCV 2024) · better on 1 of 1 shared result
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 0.87)
- ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation (NeurIPS 2024 · similarity 0.86)
- PMNI: Pose-free Multi-view Normal Integration for Reflective and Textureless Surface Reconstruction (CVPR 2025 · similarity 0.86)
- NeuSurf: On-Surface Priors for Neural Surface Reconstruction from Sparse Input Views (AAAI 2024 · similarity 0.85)
- Improving Robustness for Joint Optimization of Camera Pose and Decomposed Low-Rank Tensorial Radiance Fields (AAAI 2024 · similarity 0.85)
- Flow-Corrected Shape Optimization: Taming Manifold Drift in High-Dimensional 3D Models (NeurIPS 2026 · similarity 0.85)
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.
