DiffHairCard: Auto Hair Card Extraction with Differentiable Rendering

2025-05-24

AI Summary

This work addresses the challenge of efficiently converting strand-based hair models into lightweight hair card representations while preserving visual fidelity under strict constraints on card count and texture budget. The proposed method introduces: (1) a differentiable hair representation based on texture-space projected splines; (2) a novel two-stage co-optimization framework that jointly refines card geometry and texture; and (3) support for hair cap modeling, inter-card continuity enforcement, and level-of-detail (LoD)-aware texture sharing. End-to-end optimization is achieved via differentiable rendering, 2D spline parameterization, and hierarchical clustering. Experimental results demonstrate that the generated hair cards faithfully reproduce the appearance of original strand models across diverse hairstyles—including straight, wavy, curly, and tightly coiled hair—while significantly reducing memory footprint and rendering overhead.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Mixed Discrete/Continuous Search
• Constraint Satisfaction and Optimization: Mixed Discrete/Continuous Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Hair cards remain a widely used representation for hair modeling in real-time applications, offering a practical trade-off between visual fidelity, memory usage, and performance. However, generating high-quality hair card models remains a challenging and labor-intensive task. This work presents an automated pipeline for converting strand-based hair models into hair card models with a limited number of cards and textures while preserving the hairstyle appearance. Our key idea is a novel differentiable representation where each strand is encoded as a projected 2D spline in the texture space, which enables efficient optimization with differentiable rendering and structured results respecting the hair geometry. Based on this representation, we develop a novel algorithm pipeline, where we first cluster hair strands into initial hair cards and project the strands into the texture space. We then conduct a two-stage optimization where our first stage optimizes the texture and geometry of each hair card separately, and after texture reduction, our second stage conducts joint optimization of all the cards for fine-tuning. Put together, our method is evaluated on a wide range of hairstyles, including straight, wavy, curly, and coily hairs. To better capture the appearance of short or coily hair, we additionally support hair cap and cross-card. Furthermore, our framework supports seamless LoD transitions via texture sharing, balancing texture memory efficiency and visual quality.

Problem

• Automates conversion of strand-based to hair card models
• Optimizes hair card texture and geometry efficiently
• Supports diverse hairstyles and LoD transitions

Innovation

• Differentiable rendering for hair card optimization
• Two-stage texture and geometry optimization
• Texture sharing for seamless LoD transitions