Research map / 3D vision / Face and shape
LoG3D: Ultra-High-Resolution 3D Shape Modeling via Local-to-Global Partitioning
Xinran Yang, Shuichang Lai, Jiangjing Lyu, Hongjie Li, Bowen Pan, Yuanqi Li, Jie Guo, Zhengkang Zhou, Yanwen Guo
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 UBlock partitioning with local 3D convolutions instead of standard voxel down/up-sampling (ablation) | LoG-VAE · Toys4k (1024) | NMSE: 2.13 -> 0.34 | Improved |
| add UBlock partitioning with local 3D convolutions (ablation) | LoG-VAE · Toys4k (1024) | SNE: 3.74 -> 1.13 | Improved |
| add global sparse transformers (ablation) | LoG-VAE · Toys4k (1024) | NMSE: 0.96 -> 0.34 | Improved |
| add global sparse transformers (ablation) | LoG-VAE · Toys4k (1024) | SNE: 2.92 -> 1.13 | Improved |
| add Pad-Average strategy (ablation) | LoG-VAE · Toys4k (1024) | NMSE: 0.58 -> 0.34 | Improved |
| add Pad-Average strategy (ablation) | LoG-VAE · Toys4k (1024) | SNE: 1.90 -> 1.13 | Improved |
| add Pad-Average strategy (ablation) | LoG-VAE · Toys4k (1024) | F1(0.001): 42.55 -> 42.85 | No clear effect |
| increase padding value from 0 to 2 (ablation) | Pad-Average · Toys4k (1024) | NMSE: 0.58 -> 0.34 | Improved |
| increase padding value from 2 to 3 (ablation) | Pad-Average · Toys4k (1024) | NMSE: 0.34 -> 0.33 | No clear effect |
| scale resolution from 1024 to 2048 with same model | LoG-VAE · Toys4k | SNE: 1.13 -> 0.85 | Improved |
| scale resolution from 1024 to 2048 | LoG-VAE · Toys4k | CD: 0.06 -> 0.06 | No clear effect |
Closest papers by meaning
- Sparc3D: Sparse Representation and Construction for High-Resolution 3D Shapes Modeling (NeurIPS 2025 · similarity 0.89)
- A Lightweight UDF Learning Framework for 3D Reconstruction Based on Local Shape Functions (CVPR 2025 · similarity 0.88)
- Direct3D-S2: Gigascale 3D Generation Made Easy with Spatial Sparse Attention (NeurIPS 2025 · similarity 0.88)
- SILSA: Sliding-Window Slice Latents for Topology-Preserving High-Resolution 3D Generation (NeurIPS 2026 · similarity 0.88)
- Details Enhancement in Unsigned Distance Field Learning for High-fidelity 3D Surface Reconstruction (AAAI 2025 · similarity 0.87)
- Dora: Sampling and Benchmarking for 3D Shape Variational Auto-Encoders (CVPR 2025 · 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.
