Research map / 3D vision / Face and shape
StdGEN++: A Comprehensive System for Semantic-Decomposed 3D Character Generation
Yuze He, Yanning Zhou, Wang Zhao, Jingwen Ye, Zhongkai Wu, Ran Yi, Yong‐Jin Liu
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 coarse-to-fine proposal scheme for high-resolution extraction (ablation) | StdGEN hair layer · Hair layer ablation | F1@0.5%: 0.6416 to 0.6995 | Improved |
| add facial LoRA branch (dual-branch S-LRM) (ablation) | StdGEN++ hair layer · Hair layer ablation | F1@0.5%: 0.6995 to 0.7245 | Improved |
| add coarse-to-fine proposal scheme (ablation) | StdGEN hair layer · Hair layer ablation | Chamfer distance: 0.0506 to 0.0421 | Improved |
| add facial branch on top of coarse-to-fine (ablation) | StdGEN++ hair layer · Hair layer ablation | Chamfer distance: 0.0421 to 0.0363 | Improved |
| upgrade StdGEN to StdGEN++ (resolution, dual-branch, hole-filling) | StdGEN · Layered geometry (Hair) | F1@0.5%: 0.642 to 0.725 | Improved |
| upgrade StdGEN to StdGEN++ | StdGEN · Layered geometry (Whole) | F1@0.5%: 0.594 to 0.628 | Improved |
| upgrade StdGEN to StdGEN++ | StdGEN · Layered geometry (Body) | Chamfer distance: 0.0404 to 0.0357 | Improved |
| upgrade StdGEN to StdGEN++ | StdGEN · Layered geometry (Cloth) | Voxel IoU: 0.4345 to 0.4682 | Improved |
| upgrade StdGEN to StdGEN++ | StdGEN · A-pose input (Anime3D++ test), 3D | LPIPS: 0.066 to 0.064 | No clear effect |
| upgrade StdGEN to StdGEN++ | StdGEN · Arbitrary-pose input (Anime3D++ test), 3D | LPIPS: 0.084 to 0.084 | No clear effect |
| apply test-time multi-layer refinement to cloth and hair (ablation) | StdGEN++ · Layered geometry (Hair) | Chamfer distance | Hurt |
| raise grid resolution from 100x100x150 to 256x256x384 via sparse evaluation | StdGEN S-LRM · Layered geometry | qualitative | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- StdGEN — StdGEN: Semantic-Decomposed 3D Character Generation from Single Images (CVPR 2025) · better on 7 of 9 shared results
- Unique3D — Unique3D: High-Quality and Efficient 3D Mesh Generation from a Single Image (NeurIPS 2024) · better on 3 of 3 shared results
- Magic123 — Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priors (ICLR 2024) · 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
- StdGEN: Semantic-Decomposed 3D Character Generation from Single Images (CVPR 2025 · similarity 0.96)
- LeGO: Leveraging a Surface Deformation Network for Animatable Stylized Face Generation with One Example (CVPR 2024 · similarity 0.86)
- Text-Guided 3D Face Synthesis - From Generation to Editing (CVPR 2024 · similarity 0.86)
- Structured 3D Latents for Scalable and Versatile 3D Generation (CVPR 2025 · similarity 0.86)
- Styl3R: Instant 3D Stylized Reconstruction for Arbitrary Scenes and Styles (NeurIPS 2025 · similarity 0.86)
- Wukong's 72 Transformations: High-fidelity Textured 3D Morphing via Flow Models (NeurIPS 2025 · 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.
