DGH: Dynamic Gaussian Hair

2025-12-18

AI Summary

Dynamic hair modeling faces three core challenges: highly complex motion, severe self-occlusion, and intricate light scattering. Existing approaches—either physics-based simulation or static capture—suffer from poor generalization, high computational cost, and heavy reliance on manual parameter tuning. This paper proposes the first end-to-end, data-driven framework for dynamic strand reconstruction. Our method (1) introduces a coarse-to-fine, temporally consistent motion modeling mechanism; (2) employs strand-guided dynamic 3D Gaussian representations; and (3) establishes the first differentiable rendering-enabled Gaussian optimization pipeline for dynamic hair. Fully decoupled from physics simulation, our approach achieves strong generalization across diverse hairstyles and head motions. It significantly outperforms state-of-the-art methods in both geometric accuracy and visual fidelity, supports real-time rendering and high-fidelity animation, and integrates seamlessly into 3D Gaussian-based digital human systems.

Technology Category

• Computer Vision: 3D Computer Vision
• Humans and AI: Crowd Sourcing and Human Computation
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

The creation of photorealistic dynamic hair remains a major challenge in digital human modeling because of the complex motions, occlusions, and light scattering. Existing methods often resort to static capture and physics-based models that do not scale as they require manual parameter fine-tuning to handle the diversity of hairstyles and motions, and heavy computation to obtain high-quality appearance. In this paper, we present Dynamic Gaussian Hair (DGH), a novel framework that efficiently learns hair dynamics and appearance. We propose: (1) a coarse-to-fine model that learns temporally coherent hair motion dynamics across diverse hairstyles; (2) a strand-guided optimization module that learns a dynamic 3D Gaussian representation for hair appearance with support for differentiable rendering, enabling gradient-based learning of view-consistent appearance under motion. Unlike prior simulation-based pipelines, our approach is fully data-driven, scales with training data, and generalizes across various hairstyles and head motion sequences. Additionally, DGH can be seamlessly integrated into a 3D Gaussian avatar framework, enabling realistic, animatable hair for high-fidelity avatar representation. DGH achieves promising geometry and appearance results, providing a scalable, data-driven alternative to physics-based simulation and rendering.

Problem

• Efficiently learns dynamic hair motion and appearance
• Generalizes across diverse hairstyles and head motions
• Provides scalable data-driven alternative to physics simulation

Innovation

• Coarse-to-fine model for temporally coherent hair motion
• Strand-guided optimization for dynamic 3D Gaussian representation
• Data-driven framework generalizing across hairstyles and motions