Research map / 3D vision / Face and shape
Geo-Neus: Geometry-Consistent Neural Implicit Surfaces Learning for Multi-view Reconstruction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 view-aware SDF loss from sparse SfM points (ablation) | NeuS baseline · DTU | mean Chamfer distance: 0.87 -> 0.62 | Helped |
| add multi-view photometric consistency loss (ablation) | NeuS baseline · DTU | mean Chamfer distance: 0.87 -> 0.54 | Helped |
| add photometric loss on top of SDF loss (ablation) | Model-A · DTU | mean Chamfer distance: 0.62 -> 0.51 | Helped |
| add SDF loss on top of photometric loss (ablation) | Model-B · DTU | mean Chamfer distance: 0.54 -> 0.51 | Helped |
| locate surface points via SDF zero level set instead of depth integral for sparse point supervision (ablation) | Geo-Neus · DTU | mean Chamfer distance: 0.85 -> 0.62 | Helped |
| locate surface via SDF zero level set instead of depth integral for photometric constraint (ablation) | Geo-Neus · DTU | mean Chamfer distance: 1.08 -> 0.57 | Helped |
| add multi-view constraints on depth-integral surface (rather than SDF location) (ablation) | NeuS baseline · DTU | mean Chamfer distance: 0.87 -> 1.08 | Hurt |
| Geo-Neus instead of NeuralWarp on scan 63 | DTU scan 63 · DTU scan 63 | Chamfer distance: 0.79 -> 0.800 | No clear effect |
| add explicit SDF optimization (Geo-Neus vs NeuS) | NeuS · DTU | training time: 16 hours -> 10 hours | Helped |
Closest papers by meaning
- NeuSurf: On-Surface Priors for Neural Surface Reconstruction from Sparse Input Views (AAAI 2024 · similarity 0.92)
- SC-NeuS: Consistent Neural Surface Reconstruction from Sparse and Noisy Views (AAAI 2024 · similarity 0.90)
- Neural Signed Distance Function Inference through Splatting 3D Gaussians Pulled on Zero-Level Set (NeurIPS 2024 · similarity 0.89)
- SuperNormal: Neural Surface Reconstruction via Multi-View Normal Integration (CVPR 2024 · similarity 0.89)
- NeuRodin: A Two-stage Framework for High-Fidelity Neural Surface Reconstruction (NeurIPS 2024 · similarity 0.88)
- NC-SDF: Enhancing Indoor Scene Reconstruction Using Neural SDFs with View-Dependent Normal Compensation (CVPR 2024 · similarity 0.88)
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.