GroomCap: High-Fidelity Prior-Free Hair Capture

2024-09-01

ACM Transactions on Graphics

AI Summary

Existing hair reconstruction methods rely heavily on external priors and struggle to achieve strand-level geometric fidelity. To address this, we propose a prior-free, multi-view high-fidelity hair reconstruction framework. Our key contributions are threefold: (1) the first neural implicit volumetric hair representation that jointly encodes hair geometry and orientation; (2) a voxelized 3D orientation rendering algorithm, augmented with 2D orientation distribution supervision to enforce directional consistency; and (3) a photometrically consistent, cascaded Gaussian strand optimization strategy that jointly refines geometry and appearance. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art approaches in geometric accuracy, fine-scale detail preservation, and cross-scene generalization. The reconstructed hair models support high-quality animation and virtual human generation, enabling realistic dynamic hair synthesis without manual intervention or domain-specific priors.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Despite recent advances in multi-view hair reconstruction, achieving strand-level precision remains a significant challenge due to inherent limitations in existing capture pipelines. We introduce GroomCap , a novel multi-view hair capture method that reconstructs faithful and high-fidelity hair geometry without relying on external data priors. To address the limitations of conventional reconstruction algorithms, we propose a neural implicit representation for hair volume that encodes high-resolution 3D orientation and occupancy from input views. This implicit hair volume is trained with a new volumetric 3D orientation rendering algorithm, coupled with 2D orientation distribution supervision, to effectively prevent the loss of structural information caused by undesired orientation blending. We further propose a Gaussian-based hair optimization strategy to refine the traced hair strands with a novel chained Gaussian representation, utilizing direct photometric supervision from images. Our results demonstrate that GroomCap is able to capture high-quality hair geometries that are not only more precise and detailed than existing methods but also versatile enough for a range of applications.

Problem

• Achieving strand-level precision in hair reconstruction
• Reconstructing hair without external data priors
• Preventing structural information loss in orientation blending

Innovation

• Neural implicit representation for hair volume
• Volumetric 3D orientation rendering algorithm
• Gaussian-based hair optimization strategy