Skip to content

← ArchivePaper2022

NeuralHDHair: Automatic High-Fidelity Hair Modeling from a Single Image Using Implicit Neural Representations

Keyu Wu, Yifan Ye, Lingchen Yang, Hongbo Fu, Kun Zhou, Youyi Zheng

CVPRAcademic43 citesCFXML Deformation

Implicit neural representation (IRHairNet) hierarchically infers 3D orientation and occupancy; GrowingNet generates high-fidelity strands from a single image.

Abstract

Undoubtedly, high-fidelity 3D hair plays an indispensable role in digital humans. However, existing monocular hair modeling methods are either tricky to deploy in digital systems (e.g., due to their dependence on complex user interactions or large databases) or can produce only a coarse geometry. In this paper, we introduce NeuralHDHair, a flexible, fully automatic system for modeling high-fidelity hair from a single image. The key enablers of our system are two carefully designed neural networks: an IRHairNet (Im-plicit representation for hair using neural network) for inferring high-fidelity 3D hair geometric features (3D orientation field and 3D occupancy field) hierarchically and a GrowingNet (Growing hair strands using neural network) to efficiently generate 3D hair strands in parallel. Specifically, we perform a coarse-to-fine manner and propose a novel voxel-aligned implicit function (VIFu) to represent the global hair feature, which is further enhanced by the local details extracted from a hair luminance map. To improve the efficiency of a traditional hair growth algorithm, we adopt a local neural implicit function to grow strands based on the estimated 3D hair geometric features.

How to read this

Category
Method: single-image hair modeling via implicit neural representations
Contributions
  • IRHairNet, which hierarchically infers a 3D orientation field and 3D occupancy field from one image in a coarse-to-fine manner
  • A voxel-aligned implicit function (VIFu) for the global hair feature, enhanced by local details from a hair luminance map
  • GrowingNet, a local neural implicit function that grows 3D strands in parallel for efficiency
Context
Builds on learned single-view hair reconstruction such as HairNet (Zhou et al. 2018), moving from convolutional regression toward implicit-function representations of hair geometry.Builds on: HairNet: Single-View Hair Reconstruction Using Convolutional Neural Networks
Correctness
Demonstrated as a fully automatic single-image system aiming for high-fidelity strands without heavy user interaction or large databases, but single-view inference of occluded interior hair remains an inherently ill-posed reconstruction problem to keep in mind.
Clarity
Reasonably accessible; a first pass conveys the coarse-to-fine pipeline, a second pass is needed for the VIFu and GrowingNet formulations.
How to read it
Focus first on how the orientation and occupancy fields are defined and on the VIFu idea; do a second pass on GrowingNet if you care about strand-synthesis efficiency.

Builds on

Built upon by

Nothing yet.

Related work

Keywords

This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →