Research map / 3D vision / NeRF and radiance fields
DPA-Net: Structured 3D Abstraction from Sparse Views via Differentiable Primitive Assembly
Fenggen Yu, Yiming Qian, Xu Zhang, Francisca Gil-Ureta, Brian Jackson, Eric Bennett, Hao Zhang
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 test-time adaptation (TTA) (ablation) | DPA-Net · ShapeNet chair | CD: 1.22 -> 0.79 | Improved |
| add silhouette-aware adaptive pixel sampling (ablation) | DPA-Net · ShapeNet chair | CD: 1.11 -> 0.79 | Improved |
| add overlapping loss (ablation) | DPA-Net · ShapeNet chair | #P (primitives): 61.23 -> 42.38 | Improved |
| add primitive dropout (ablation) | DPA-Net · ShapeNet chair | #P (primitives): 59.64 -> 42.38 | Improved |
| add primitive dropout (ablation) | DPA-Net · ShapeNet chair | CD: 0.81 -> 0.79 | No clear effect |
| structured primitive assembly instead of unstructured pixelNeRF mesh, multi-category training | DPA-Net · ShapeNet category-agnostic | CD: 3.67 -> 1.47 | Improved |
| DPA-Net versus pixelNeRF on chair category | DPA-Net · ShapeNet chair | CD: 0.50 -> 0.79 | Hurt |
| DPA-Net versus EMS on normal consistency in category-agnostic setting | DPA-Net · ShapeNet category-agnostic | NC: 0.75 -> 0.70 | Hurt |
| DPA-Net versus DBW on DTU | DPA-Net · DTU 6 scenes | mean CD: 4.12 -> 3.41, parts 3.83 -> 7.33 | Mixed |
Closest papers by meaning
- Soft Rasterizer: A Differentiable Renderer for Image-Based 3D Reasoning (ICCV 2019 · similarity 0.88)
- LinPrim: Linear Primitives for Differentiable Volumetric Rendering (NeurIPS 2025 · similarity 0.87)
- RenderNet: A deep convolutional network for differentiable rendering from 3D shapes (NeurIPS 2018 · similarity 0.87)
- A General Differentiable Mesh Renderer for Image-Based 3D Reasoning (TPAMI 2020 · similarity 0.87)
- NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis (ECCV 2020 · similarity 0.87)
- DNGaussian: Optimizing Sparse-View 3D Gaussian Radiance Fields with Global-Local Depth Normalization (CVPR 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.
