Research map / 3D vision / Face and shape
StrandHead: Text to Hair-Disentangled 3D Head Avatars Using Human-Centric Priors
Xiaokun Sun, Zeyu Cai, Ying Tai, Jian Yang, Zhenyu 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 |
|---|---|---|---|
| use StrandHead instead of HAAR for hair generation | text-to-hair · Text-to-hair generation (30 prompts) | BLIP2-VQA: 0.2000 -> 0.9000 | Improved |
| use StrandHead instead of HAAR for hair generation | text-to-hair · Text-to-hair generation (30 prompts) | FID: 335.07 -> 201.19 | Improved |
| use StrandHead instead of TECA for head generation | text-to-head · Text-to-head generation (30 prompts) | generation quality preference (%): 34.33 -> 53.67 | Improved |
| StrandHead versus LucidDreamer on hair | text-to-hair · Text-to-hair generation (30 prompts) | BLIP2-VQA: 0.9333 -> 0.9000 | Hurt |
| add human-specific 2D generative priors (vs no or general 2D diffusion supervision) (ablation) | StrandHead hair optimization · hair generation (qualitative) | hair realism | Improved |
| add orientation consistency loss L_ori (ablation) | StrandHead hair geometry · hair generation (qualitative) | hair naturalness | Improved |
| add curvature regularization L_cur (ablation) | StrandHead hair geometry · hair generation (qualitative) | curliness fidelity | Improved |
| add differentiable prismatization instead of NeuralHaircut quad mesh (ablation) | StrandHead hair optimization · hair generation (qualitative) | optimization stability | Improved |
| add FLAME-evolving prior loss (vs none or non-evolving prior) (ablation) | StrandHead bald head generation · bald head generation (qualitative) | head geometry | Improved |
| add geometry-aware losses L_bbox, L_face, L_colli (ablation) | StrandHead hair geometry · hair generation (qualitative) | geometric rationality | Improved |
| add strand-aware texture field (UV coordinates plus orientation) (ablation) | StrandHead hair texture · hair generation (qualitative) | texture realism | Improved |
| increase number of lateral edges K of the prism (ablation) | differentiable prismatization · hair generation (qualitative) | output quality | No clear effect |
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.
- HumanNorm — HumanNorm: Learning Normal Diffusion Model for High-quality and Realistic 3D Human Generation (CVPR 2024) · better on 4 of 4 shared results
- LucidDreamer — LucidDreamer: Towards High-Fidelity Text-to-3D Generation via Interval Score Matching (CVPR 2024) · better on 3 of 4 shared results
- HeadStudio — HeadStudio: Text to Animatable Head Avatars with 3D Gaussian Splatting (ECCV 2024) · better on 4 of 4 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
- Text-Conditioned Generative Model of 3D Strand-based Human Hairstyles (CVPR 2024 · similarity 0.90)
- Doubly Hierarchical Geometric Representations for Strand-based Human Hairstyle Generation (NeurIPS 2024 · similarity 0.88)
- Perm: A Parametric Representation for Multi-Style 3D Hair Modeling (ICLR 2025 · similarity 0.88)
- SimAvatar: Simulation-Ready Avatars with Layered Hair and Clothing (CVPR 2025 · similarity 0.88)
- LUCAS: Layered Universal Codec Avatars (CVPR 2025 · similarity 0.88)
- DiffLocks: Generating 3D Hair from a Single Image using Diffusion Models (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.
