Research map / 3D vision / Face and shape
"NeuSDFusion: A Spatial-Aware Generative Model for 3D Shape Completion, Reconstruction, and Generation"
Ruikai Cui, Weizhe Liu, Weixuan Sun, Senbo Wang, Taizhang Shang, Yang Li, Xibin Song, Han Yan, ZHENNAN WU, Shenzhou Chen, HONGDONG LI, Pan Ji
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 NeuSDF tri-plane latent diffusion instead of 3DQD TSDF representation | NeuSDFusion · ShapeNet Airplane unconditional | 1-NNA CD: 56.29 -> 52.33 | Improved |
| NeuSDFusion vs LION on Chair 1-NNA CD | NeuSDFusion · ShapeNet Chair unconditional | 1-NNA CD: 53.70 -> 51.95 | Improved |
| NeuSDFusion vs 3DQD on shape completion quality (MMD, bottom half) | NeuSDFusion · ShapeNet shape completion (bottom half) | MMD (x10^2): 2.93 -> 2.29 | Improved |
| NeuSDFusion vs 3DQD on completion diversity (TMD, octant) | NeuSDFusion · ShapeNet shape completion (octant) | TMD (x10^2): 9.60 -> 8.32 | Hurt |
| NeuSDFusion vs SDFusion on Pix3D single-view reconstruction | NeuSDFusion · Pix3D single-view reconstruction | Chamfer Distance: 1.85 -> 0.92 | Improved |
| NeuSDFusion vs 3DQD on text-guided FPD | NeuSDFusion · ShapeGlot text-guided generation | FPD: 59.00 -> 55.01 | Improved |
| NeuSDFusion vs 3DQD on text-guided PMMD | NeuSDFusion · ShapeGlot text-guided generation | PMMD (x10^2): 1.49 -> 1.49 | No clear effect |
| use spatial-aware transformer autoencoder instead of roll-out or channel-wise concatenation CNN autoencoder (ablation) | NeuSDFusion autoencoder · ShapeNet Chair reconstruction | reconstruction quality (qualitative) | Improved |
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.
- SDFusion — SDFusion: Multimodal 3D Shape Completion, Reconstruction, and Generation (CVPR 2023) · 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
- Diffusion-Based Signed Distance Fields for 3D Shape Generation (CVPR 2023 · similarity 0.90)
- Unsupervised Primitive Discovery for Improved 3D Generative Modeling (CVPR 2019 · similarity 0.90)
- High-Fidelity 3D Textured Shapes Generation by Sparse Encoding and Adversarial Decoding (ECCV 2024 · similarity 0.89)
- GENA3D: Generative Amodal 3D Modeling by Bridging 2D Priors and 3D Coherence (ECCV 2026 · similarity 0.89)
- Make-A-Shape: a Ten-Million-scale 3D Shape Model (ICML 2024 · similarity 0.88)
- SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation (NeurIPS 2026 · 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.
