Geometry in Style: 3D Stylization via Surface Normal Deformation

2025-03-29

AI Summary

Existing 3D mesh stylization methods struggle to simultaneously preserve shape identity and ensure expressive geometric control: weak deformations (e.g., normal maps) lack fidelity, while strong deformations often introduce artifacts or compromise source topology and semantics. This paper proposes a target-normal-based triangular mesh stylization framework, the first to employ vertex-neighborhood target normals as deformation parameters—enabling fine-grained, interpretable geometric control. Its core innovation is a differentiable As-Rigid-As-Possible (dARAP) layer, enabling end-to-end neural network optimization jointly with rigidity constraints. Integrated with vision-language guidance—specifically, a text-to-image model’s visual loss—it achieves text-driven, artifact-free, high-fidelity 3D mesh stylization. Experiments demonstrate that our method strictly preserves the source shape’s identity while significantly enhancing stylistic detail and geometric consistency across diverse inputs.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Sentiment Analysis, Stylistic Analysis, and Argument Mining
• Machine Learning: Deep Neural Architectures and Foundation Models

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

We present Geometry in Style, a new method for identity-preserving mesh stylization. Existing techniques either adhere to the original shape through overly restrictive deformations such as bump maps or significantly modify the input shape using expressive deformations that may introduce artifacts or alter the identity of the source shape. In contrast, we represent a deformation of a triangle mesh as a target normal vector for each vertex neighborhood. The deformations we recover from target normals are expressive enough to enable detailed stylizations yet restrictive enough to preserve the shape's identity. We achieve such deformations using our novel differentiable As-Rigid-As-Possible (dARAP) layer, a neural-network-ready adaptation of the classical ARAP algorithm which we use to solve for per-vertex rotations and deformed vertices. As a differentiable layer, dARAP is paired with a visual loss from a text-to-image model to drive deformations toward style prompts, altogether giving us Geometry in Style. Our project page is at https://threedle.github.io/geometry-in-style.

Problem

• Identity-preserving 3D mesh stylization with expressive deformations
• Balancing shape preservation and detailed stylization via normal vectors
• Differentiable ARAP layer for neural-network-driven deformation and style adaptation

Innovation

• Uses target normal vectors for mesh deformation
• Employs differentiable ARAP layer for shape preservation
• Integrates text-to-image model for style guidance