用单视频生成可重打光的高保真3D人脸，效果更真实。
HRAvatar: High-Quality and Relightable Gaussian Head Avatar
- 端到端优化减少追踪误差，用可学习混合形状捕捉表情变化。
- 在真实光照下重建效果优于现有方法，支持动态光影变化。
- 适合做虚拟形象、影视特效的开发者或研究人员参考。
从单视角视频重建可驱动且高质量的3D人脸头像，尤其是实现真实重打光，具有重要价值。然而，单视图输入信息有限，叠加复杂的头部姿态与面部运动，使该任务极具挑战。现有方法虽结合3D高斯泼溅（3DGS）与参数化头模型实现实时性能，但因面部追踪不准及变形模型表达能力有限，导致头部质量下降，且无法在新光照条件下产生真实效果。为此，我们提出HRAvatar，一种基于3DGS的方法，可重建高保真、可重打光的3D头像。通过端到端优化降低追踪误差，并采用可学习混合形状（learnable blendshapes）和可学习线性混合蒙皮（learnable linear blend skinning）更好捕捉个体面部形变。此外，将头像外观分解为多个物理属性，并引入基于物理的着色模型以模拟环境光照。大量实验表明，HRAvatar不仅重建出更高品质的人脸，还能在不同光照条件下呈现逼真的视觉效果。
原文摘要 · Abstract (English)
Reconstructing animatable and high-quality 3D head avatars from monocular videos, especially with realistic relighting, is a valuable task. However, the limited information from single-view input, combined with the complex head poses and facial movements, makes this challenging. Previous methods achieve real-time performance by combining 3D Gaussian Splatting with a parametric head model, but the resulting head quality suffers from inaccurate face tracking and limited expressiveness of the deformation model. These methods also fail to produce realistic effects under novel lighting conditions. To address these issues, we propose HRAvatar, a 3DGS-based method that reconstructs high-fidelity, relightable 3D head avatars. HRAvatar reduces tracking errors through end-to-end optimization and better captures individual facial deformations using learnable blendshapes and learnable linear blend skinning. Additionally, it decomposes head appearance into several physical properties and incorporates physically-based shading to account for environmental lighting. Extensive experiments demonstrate that HRAvatar not only reconstructs superior-quality heads but also achieves realistic visual effects under varying lighting conditions.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。