Research map / 3D vision / Face and shape
Im2Haircut: Single-view Strand-based Hair Reconstruction for Human Avatars
Vanessa Sklyarova, Egor Zakharov, Malte Prinzler, Giorgio Becherini, Michael J. Black, Justus Thies
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 PCA component loss in coarse stage (ablation) | coarse branch · Real images (33) | mask: 0.598 -> 0.517 | Improved |
| add curvature loss in coarse stage (ablation) | coarse branch · Real images (33) | mask: 0.545 -> 0.517 | Improved |
| add curvature loss in coarse stage (ablation) | coarse branch · Synthetic hairstyles (50) | chamfer_pts: 0.00022 -> 0.00026 | Hurt |
| add depth map as input (ablation) | coarse branch · Real images (33) | mask: 0.548 -> 0.517 | Improved |
| add direction loss in coarse stage (ablation) | coarse branch · Real images (33) | mask: 0.553 -> 0.517 | Improved |
| add real-data mixing to prior training (ablation) | fine-stage prior · Real images (33) | mask: 0.547 -> 0.405 | Improved |
| add real-data mixing to prior training (ablation) | fine-stage prior · Synthetic hairstyles (50) | chamfer_angle: 0.128 -> 0.143 | Hurt |
| add coarse-to-fine training instead of single-stage 64 components (ablation) | prior model · Real images (33) | mask: 0.578 -> 0.405 | Improved |
| add mixing of real data rather than rendering loss on synthetic data only (ablation) | prior model · Real images (33) | mask: 0.519 -> 0.405 | Improved |
| add penetration loss to prior training (ablation) | prior model · Real images (33) | mask: 0.412 -> 0.405; chamfer_angle 0.139 -> 0.143 | Mixed |
| add silhouette loss to prior training (ablation) | prior model · Real images (33) | mask: 0.571 -> 0.405 | Improved |
| add direction loss to prior training (ablation) | prior model · Real images (33) | L_undir: 0.723 -> 0.695 | Improved |
Later papers that compared with this method
1 later accepted paper reports Im2Haircut as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- HairOrbit: Multi-view Aware 3D Hair Modeling from Single Portraits (ECCV 2026 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- PERM — Perm: A Parametric Representation for Multi-Style 3D Hair Modeling (ICLR 2025) · better on 2 of 2 shared results
- NeuralHDHair — NeuralHDHair: Automatic High-Fidelity Hair Modeling From a Single Image Using Implicit Neural Representations (CVPR 2022) · 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
- MonoHair: High-Fidelity Hair Modeling from a Monocular Video (CVPR 2024 · similarity 0.93)
- Dr.Hair: Reconstructing Scalp-Connected Hair Strands without Pre-Training via Differentiable Rendering of Line Segments (CVPR 2024 · similarity 0.92)
- DiffLocks: Generating 3D Hair from a Single Image using Diffusion Models (CVPR 2025 · similarity 0.92)
- Perm: A Parametric Representation for Multi-Style 3D Hair Modeling (ICLR 2025 · similarity 0.91)
- Text-Conditioned Generative Model of 3D Strand-based Human Hairstyles (CVPR 2024 · similarity 0.91)
- Doubly Hierarchical Geometric Representations for Strand-based Human Hairstyle Generation (NeurIPS 2024 · similarity 0.90)
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.
