Perm: A Parametric Representation for Multi-Style 3D Hair Modeling

2024-07-28

arXiv.org

AI Summary

To address the lack of flexibility and controllability in 3D human hair modeling and editing, this paper proposes Perm—a learnable parametric representation for multi-style hairstyles. Our method tackles three core challenges: (1) We introduce a novel strand representation in the frequency domain, coupled with frequency-domain PCA decomposition, to disentangle global hair structure from local curl patterns; (2) We design a geometric decomposition mechanism that separates guide texture (structural backbone) from residual texture (fine-scale details), enabling hierarchical modeling of geometry; (3) We formulate hair styling as a layered generative process. Evaluated on single-view 3D hair reconstruction, interactive editing, and hairstyle-conditional image generation, Perm achieves state-of-the-art performance across all tasks. It demonstrates strong generalization capability and seamless cross-task deployment, establishing a unified, controllable framework for diverse hair-related applications.

Technology Category

• Computer Vision: Representation Learning for Vision
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

We present Perm, a learned parametric representation of human 3D hair designed to facilitate various hair-related applications. Unlike previous work that jointly models the global hair structure and local curl patterns, we propose to disentangle them using a PCA-based strand representation in the frequency domain, thereby allowing more precise editing and output control. Specifically, we leverage our strand representation to fit and decompose hair geometry textures into low- to high-frequency hair structures, termed guide textures and residual textures, respectively. These decomposed textures are later parameterized with different generative models, emulating common stages in the hair grooming process. We conduct extensive experiments to validate the architecture design of Perm, and finally deploy the trained model as a generic prior to solve task-agnostic problems, further showcasing its flexibility and superiority in tasks such as single-view hair reconstruction, hairstyle editing, and hair-conditioned image generation. More details can be found on our project page: https://cs.yale.edu/homes/che/projects/perm/.

Problem

• Disentangle global and local 3D hair structures
• Parameterize hair textures using generative models
• Solve tasks like hair reconstruction and editing

Innovation

• PCA-based strand representation
• Decompose hair geometry textures
• Parameterized generative models