Research map / 3D vision / Face and shape
NeuDA: Neural Deformable Anchor for High-Fidelity Implicit Surface Reconstruction
Bowen Cai, Jinchi Huang, Rongfei Jia, Chengfei Lv, Huan Fu
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 hierarchical deformable anchors and the NeuDA representation instead of the strongest reported baseline | surface reconstruction · DTU w/ mask | mean CD: 0.72 -> 0.67 | Improved |
| use hierarchical deformable anchors and the NeuDA representation instead of the strongest reported baseline | surface reconstruction · DTU w/o mask | mean CD: 0.68 -> 0.65 | Improved |
| use NeuDA instead of Instant-NeuS | surface rendering · DTU train | mean PSNR: 31.33 -> 30.90 | Hurt |
| add a multi-level grid rather than using a single-level feature grid (ablation) | NeuDA representation · DTU w/ mask | mean CD: 1.25 -> 0.74 | Improved |
| replace fixed feature vectors with fixed 3D anchors (ablation) | multi-level grid · DTU w/ mask | mean CD: 0.74 -> 0.75 | No clear effect |
| optimize the anchor points (ablation) | multi-level anchor grid · DTU w/ mask | mean CD: 0.75 -> 0.67 | Improved |
| replace multi-level feature vectors with optimized deformable anchors (ablation) | multi-level grid · DTU w/ mask | mean CD: 0.74 -> 0.67 | Improved |
| add normal regularization (ablation) | NeuDA · DTU w/o mask | mean CD: 0.67 -> 0.65 | Improved |
| use hierarchical positional encoding instead of standard multi-level positional encoding (ablation) | NeuDA · DTU | mean CD: 0.67 -> 0.67 | No clear effect |
| increase the number of grid levels, with performance improving initially but slightly worsening at 10 levels (ablation) | NeuDA · DTU | Chamfer Distance | Mixed |
Later papers that compared with this method
1 later accepted paper reports NeuDA as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Rethinking Directional Parameterization in Neural Implicit Surface Reconstruction (ECCV 2024 · better on 4 of 4 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- NeuS — NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction (NeurIPS 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
- NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction (NeurIPS 2021 · similarity 0.89)
- Voxurf: Voxel-based Efficient and Accurate Neural Surface Reconstruction (ICLR 2023 · similarity 0.89)
- HF-NeuS: Improved Surface Reconstruction Using High-Frequency Details (NeurIPS 2022 · similarity 0.89)
- NeuSurf: On-Surface Priors for Neural Surface Reconstruction from Sparse Input Views (AAAI 2024 · similarity 0.87)
- Implicit-ARAP: Efficient Handle-Guided Neural Field Deformation via Local Patch Meshing (NeurIPS 2025 · similarity 0.86)
- SC-NeuS: Consistent Neural Surface Reconstruction from Sparse and Noisy Views (AAAI 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.
