Research map / 3D vision / Face and shape
A Probability-guided Sampler for Neural Implicit Surface Rendering
Gonçalo José Dias Pais, Valter André Piedade, Moitreya Chatterjee, Marcus Greiff, Pedro Miraldo
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 probability-guided sampler and surface losses to Neuralangelo | Neuralangelo · DTU (unmasked, mean) | PSNR: 33.84 -> 34.41 | Improved |
| add probability-guided sampler and surface losses to NeuS | NeuS · DTU (unmasked, mean) | PSNR: 31.74 -> 31.78 | No clear effect |
| add sampler to NeuS | NeuS · DTU (mean) | Chamfer distance (mm): 1.30 -> 1.2 | Improved |
| add sampler to Neuralangelo | Neuralangelo · DTU (mean) | Chamfer distance (mm): 0.61 -> 0.6 | Improved |
| add sampler to Neuralangelo on large-scale TNT scenes | Neuralangelo · TNT | PSNR: 25.99 -> 25.65 | Hurt |
| add sampler to Neuralangelo on TNT | Neuralangelo · TNT | F1: 0.58 -> 0.59 | Improved |
| add surface reconstruction losses (weight 500) (ablation) | Neuralangelo + sampler · DTU subset (24, 65, 97, 122) | Chamfer distance (mm): 0.82 -> 0.57 | Improved |
| add surface reconstruction losses (weight 500) (ablation) | Neuralangelo + sampler · DTU subset (24, 65, 97, 122) | PSNR: 35.44 -> 35.24 | Hurt |
| add view dependency (occlusion-aware weighting) to the sampling PDF (ablation) | Neuralangelo + sampler · DTU subset (24, 65, 97, 122) | Chamfer distance (mm): 1.04 -> 0.57 | Improved |
| add extra ray samples near the sampled surface depth (ablation) | Neuralangelo + sampler · DTU subset (24, 65, 97, 122) | PSNR: 34.82 -> 35.24 | Improved |
| update s adaptively instead of using a constant value (ablation) | Neuralangelo + sampler · DTU subset (24, 65, 97, 122) | Chamfer distance (mm): 0.58 -> 0.57 | No clear effect |
| raise surface loss weight from 10 to 500 (ablation) | Neuralangelo + sampler · DTU subset (24, 65, 97, 122) | Chamfer distance (mm): 0.85 -> 0.57 | 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.
- NeuS — NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction (NeurIPS 2021) · better on 3 of 3 shared results
- HF-NeuS — HF-NeuS: Improved Surface Reconstruction Using High-Frequency Details (NeurIPS 2022) · 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
- NeXT: Towards High Quality Neural Radiance Fields via Multi-Skip Transformer (ECCV 2022 · similarity 0.90)
- UNISURF: Unifying Neural Implicit Surfaces and Radiance Fields for Multi-View Reconstruction (ICCV 2021 · similarity 0.90)
- Neural Radiance Fields From Sparse RGB-D Images for High-Quality View Synthesis (TPAMI 2022 · similarity 0.90)
- Differentiable Surface Rendering via Non-Differentiable Sampling (ICCV 2021 · similarity 0.90)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (CVPR 2025 · similarity 0.90)
- Rad-NeRF: Ray-decoupled Training of Neural Radiance Field (NeurIPS 2024 · similarity 0.89)
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.
