Research map / 3D vision / Face and shape
Geometry Distributions
Biao Zhang, Jing Ren, Peter Wonka
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 proposed network instead of hashing grid (ablation) | Geometry Distributions · Loong shape | Chamfer x10^3: 4.133 -> 2.140 | Improved |
| use proposed network instead of DeepSDF-style MLP (ablation) | Geometry Distributions · Loong shape | Chamfer x10^3: 2.647 -> 2.140 | Improved |
| increase training set size from 10^5 to approx. infinity (resampling) (ablation) | Geometry Distributions · jellyfish shape | Chamfer x10^3: 4.696 -> 2.916 | Improved |
| increase training set from 10^7 to 10^8 samples (ablation) | Geometry Distributions · jellyfish shape | Chamfer x10^3: 2.936 -> 2.917 | No clear effect |
| increase sampling steps from 8 to 64 (ablation) | Geometry Distributions · Archimedes shape | Chamfer x10^3: 9.803 -> 2.780 | Improved |
| increase sampling steps from 16 to 64 (ablation) | Geometry Distributions · Archimedes shape | Chamfer x10^3: 2.814 -> 2.780 | No clear effect |
| increase network blocks from 2 to 10 (ablation) | Geometry Distributions · lamp shape | Chamfer x10^3: 3.784 -> 3.544 | Improved |
| increase network blocks from 4 to 10 (ablation) | Geometry Distributions · lamp shape | Chamfer x10^3: 3.565 -> 3.544 | No clear effect |
| use Gaussian instead of uniform initial noise (ablation) | Geometry Distributions · lion shape | Chamfer x10^3: 3.734 -> 3.246 | Improved |
| use Gaussian instead of uniform initial noise (ablation) | Geometry Distributions · lamp shape | Chamfer x10^3: 3.574 -> 3.538 | Improved |
| increase inversion steps from 4 to 64 (ablation) | Geometry Distributions · mouse shape | inversion MSE: 6.88e-1 -> 1.76e-6 | Improved |
| model surface as distribution instead of vector-field based method | vector field methods · uniform surface sampling | Chamfer x10^3 | Improved |
Closest papers by meaning
- Geometry Image Diffusion: Fast and Data-Efficient Text-to-3D with Image-Based Surface Representation (ICLR 2025 · similarity 0.89)
- Neural Convolutional Surfaces (CVPR 2022 · similarity 0.89)
- Surface Networks (CVPR 2018 · similarity 0.88)
- LaGeM: A Large Geometry Model for 3D Representation Learning and Diffusion (ICLR 2025 · similarity 0.87)
- Geometry Awakening: Cross-Geometry Learning Exhibits Superiority over Individual Structures (NeurIPS 2024 · similarity 0.87)
- Sin3DM: Learning a Diffusion Model from a Single 3D Textured Shape (ICLR 2024 · similarity 0.87)
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.
