Reconstructing strand-level 3D hair from a single-view image is highly challenging, especially when preserving consistent and realistic attributes in unseen regions. Existing methods rely on limited frontal-view cues and small-scale/style-restricted synthetic data, often failing to produce satisfactory results in invisible regions. In this work, we propose a novel framework that leverages the strong 3D priors of video generation models to transform single-view hair reconstruction into a calibrated multi-view reconstruction task. To balance reconstruction quality and efficiency for the reformulated multi-view task, we further introduce a neural orientation extractor trained on sparse real-image annotations for better full-view orientation estimation. In addition, we design a two-stage strand-growing algorithm based on a hybrid implicit field to synthesize the 3D strand curves with fine-grained details at a relatively fast speed. Extensive experiments demonstrate that our method achieves state-of-the-art performance on single-view 3D hair strand reconstruction on a diverse range of hair portraits in both visible and invisible regions.
Figures & tables
Figure 1: We propose a novel framework for strand-level single-view 3D hair reconstruction. Given a frontal-view portrait (a), we first synthesize corresponding calibrated multi-view images (b) on a camera orbit, then reconstruct multi-view aware 3D hair strands (c). Note that the left view in (c) is rendered with about 10k strands to better visualize the geometry, while the other 3 views are rendered with 100k.
Figure 2: Overview of HairOrbit . Given a single portrait, HairOrbit converts single-view 3D hair reconstruction into a multi-view task.
nastya
ksyusha
jenya
shortCurly
whiteCurly
midWavy
midCurly
Overall
#Views
62
66
70
58
42
33
64
395
Gabor
11.88
16.56
18.88
18.87
24.56
11.31
10.79
16.05
HairStep
6.28
10.98
13.16
13.68
21.57
6.32
5.54
10.88
Ours
1.79
4.97
6.09
6.72
13.50
2.15
2.12
5.13
Table 1: Quantitative comparisons of full-view orientation extraction (mean angle error, lower is better). #Views means the number of annotated views.
Figure 3: Qualitative comparisons of full-view orientation extraction. Results of HairStep have been converted to orientation maps.
Figure 4: Comparisons on multi-view generation.
Figure 5: Qualitative comparison on single-view 3D strands reconstruction. For every example, we show (a) the input image and the reconstructed 3D hair strands rendered in multiple views of (b) Ours, (c) Im2Haircut, (d) HairStep and (e) Difflocks.
Metric
HairStep
DiffLocks
Im2Haircut
Ours
HairSale ( ∘ ) ↓
17.38
26.50
16.24
12.83
HairRida ( % ) ↑
77.01
69.67
79.38
80.52
IoU ↑
0.639
0.593
0.757
0.847
Table 2: Quantitative comparisons on single-view 3D strands reconstruction with HairSale and IoU.
Figure 6: Qualitative comparisons of ablation study.
Method
@Input View
@Synthesized 8 Views
HairSale ( ∘ ) ↓
IoU ↑
HairSale ( ∘ ) ↓
IoU ↑
C0
16.42 (-27.98%)
0.794
13.45 (-34.90%)
0.878
C1
13.70 (-6.78%)
0.828
10.37 (-4.01%)
0.887
C2
19.10 (-48.87%)
0.680
24.14 (-142.1%)
0.658
Full
12.83
0.847
9.97
0.900
Table 3: Quantitative comparisons of ablation study. Blue values denote the performance degradation (%) compared to Full .
Strand-level hair geometry reconstruction is a fundamental problem in virtual human modeling and the digitization of hairstyles. However, existing methods still suffer from a significant trade-off between accuracy and efficiency. Implicit neural representations can capture the global hair shape but often fail to preserve fine-grained strand details, while explicit optimization-based approaches achieve high-fidelity reconstructions at the cost of heavy computation and poor scalability. To address this issue, we propose EfficientMonoHair, a fast and accurate framework that combines the implicit neural network with multi-view geometric fusion for strand-level reconstruction from monocular video. Our method introduces a fusion-patch-based multi-view optimization that reduces the number of optimization iterations for point cloud direction, as well as a novel parallel hair-growing strategy that relaxes voxel occupancy constraints, allowing large-scale strand tracing to remain stable and robust even under inaccurate or noisy orientation fields. Extensive experiments on representative real-world hairstyles demonstrate that our method can robustly reconstruct high-fidelity strand geometries with accuracy. On synthetic benchmarks, our method achieves reconstruction quality comparable to state-of-the-art methods, while improving runtime efficiency by nearly an order of magnitude.
Da Li, Dominik Engel, Deng Luo +1
King Abdullah University of Science and Technology, Saudi Arabia
The fundamental limitation of traditional strand-based modeling is not simply data scarcity, but the ill-posedness of inferring complex 3D fields from 2D imagery without structural constraints. This unconstrained regression leads to catastrophic failures in resolving both global occlusion (e.g., in ponytails) and local directionality (e.g., in curls), resulting in over-smoothed, plausible-but-incorrect geometries. To resolve this, we integrate the strong geometric priors of Large Reconstruction Models (LRMs) into the strand generation pipeline. Using the LRM mesh as a structural anchor, we employ a novel Dual Orientation AutoEncoder to lift coarse geometry into high-fidelity strands. By resolving vector field singularities through latent-space optimization and surface-guided refinement, our method effectively disentangles complex topological structures, setting a new benchmark for robustness and accuracy in hair reconstruction.
Yuefan Shen, Yican Dong, Xiufeng Huang +3
LIGHTSPEED, China · State Key Lab of CAD and CG, Zhejiang University, China · Hong Kong Baptist University, China +1
Modeling 3D hair strands is challenging due to complex hairstyle geometry, scarce high-quality training data, and incomplete or noisy input observations. Foundation models offer broader hairstyle priors, but their outputs are not directly usable for strand-based rendering and simulation. Existing approaches require costly per-instance optimization to fit hair priors to meshes from foundation models and 2D orientations from specialized predictors trained on limited hair data, restricting efficiency and generalization. We present HairFlow, a multimodal conditioned diffusion framework for efficient and controllable hair strands modeling. HairFlow conditions strand generation on two complementary representations: a coarse mesh that constrains global shape and multi-view sketches that specify local strand flow. HairFlow combines an Orientation VAE with a conditioned diffusion transformer to generate volumetric fields that are traced into explicit strands. To overcome the lack of paired sketch-strand training data, we introduce an automatic synthesis pipeline that generates geometrically aligned multi-view sketches with diverse styles representative of inference-time inputs. Experiments show that HairFlow generates plausible 3D strands that faithfully follow the input conditions across diverse hairstyles, substantially reduces inference time, and supports flexible multimodal control.