TANGLED - A 3D hairstyle generation method jointly developed by ShanghaiTech University and Huazhong University of Science and Technology.
TANGLED is a 3D hair generation method jointly developed by ShanghaiTech University, Deemos Technology, and Huazhong University of Science and Technology. It supports the generation of high-quality 3D hair strands from images of any style and perspective. TANGLED is based on three core...
What is TANGLED?
TANGLED is a 3D hairstyle generation method jointly developed by ShanghaiTech University, Deemos Technology, and Huazhong University of Science and Technology. It supports the generation of high-quality 3D hair strands from images of any style and perspective. TANGLED is based on three core steps: providing rich hairstyle samples using the diverse MultiHair dataset; capturing the topological structure of the hairstyle using cross-attention of line art features based on a diffusion framework of multi-view line art; and restoring details of complex hairstyles (such as braids) using a parametric post-processing module. TANGLED enhances the realism and diversity of hairstyles, supports the creation of culturally inclusive digital characters, and provides new application possibilities for fields such as animation and augmented reality.
Main functions of TANGLED
- Diverse hairstyle generationIt supports handling various complex hairstyles, such as braids, curls, and traditional hairstyles.
- Multi-view input supportAccepts single-view or multi-view images as input.
- Flexible input stylesIt supports multiple input styles, including photos, hand-drawn sketches, cartoons, and oil paintings, to meet the needs of different application scenarios.
- Cultural InclusivityIt pays special attention to underrepresented hair textures (such as curls and braids) and complex geometries, supporting the generation of culturally significant hairstyles.
- High-efficiency integrationThe generated 3D hairstyles can be directly integrated into existing CG workflows, such as Blender and Unreal Engine.
TANGLED's technical principles
- NeuraPressMultiHair datasetThe dataset provides 457 diverse hairstyles, labeled with 74 attributes, with a focus on complex and culturally significant styles. Diversity is enhanced through multi-view rendering and line drawing extraction, and detailed text annotations are generated using GPT-4.
- diffusion frameworkA diffusion model based on multi-view line art captures the topological structure of the hairstyle (such as hair density and parting lines). Line art features are extracted using DINOv2 and integrated into the diffusion model based on a cross-attention mechanism. Line art features from different perspectives are randomly mixed to enhance the model's adaptability to different input styles and viewpoints.
- Parameterized post-processing moduleParametric modeling and restoration techniques are designed for complex hairstyles (such as braids). The Frenet-Serret framework is used to generate the geometry of the braids, which are then naturally integrated into the hairstyle based on an attention mechanism. Laplacian smoothing technology is used to reduce high-frequency noise, ensuring the geometric coherence and visual appeal of the hairstyle.
TANGLED's project address
- Project official website:https://sites.google.com/view/tangled
- arXiv technical paper:https://arxiv.org/pdf/2502.06392v1
Application scenarios of TANGLED
- Culturally Inclusive Virtual Character CreationGenerate diverse hairstyles and support virtual character design from different cultural backgrounds.
- Animated Hairstyle DesignGenerate 3D hairstyles from sketches to improve animation production efficiency.
- Augmented Reality (AR) Hairstyle PreviewCombining AR technology allows users to virtually try on different hairstyles.
- Virtual Makeup Try-on App: Helps users preview the hairstyle effect before purchasing.
- Sketch-based 3D hair editingIt allows for quick adjustments to hairstyles based on modified sketches, making it suitable for creative design fields.