EGGHead: Efficient Generalizable Gaussian Heads from a Single Image

Computer Science > Computer Vision and Pattern Recognition arXiv:2610.26898 (cs) [Submitted on 16 Oct 2026] Title: EGGHead: Efficient Generalizable Gaussian Heads from a Single Image Authors: Pilseo Park, Andrew Hou, Michel Sarkis, Ning Bi, Yiying Tong, Xiaoming Liu.

计算机科学 > 计算机视觉与模式识别 arXiv:2610.26898 (cs) [提交于 2026 年 10 月 16 日] 标题：EGGHead：基于单张图像的高效可泛化高斯头部模型 作者：Pilseo Park, Andrew Hou, Michel Sarkis, Ning Bi, Yiying Tong, Xiaoming Liu。

Abstract: Creating photorealistic human heads typically involves trade-offs between representation power, efficiency, and generalizability. Traditional mesh-based solutions relying on 3DMMs generally sacrifice representation power but are efficient to render. On the other hand, NeRF-based methods improve representation power and fidelity while sacrificing efficiency.

摘要：创建照片级逼真的人类头部模型通常需要在表征能力、效率和泛化性之间进行权衡。依赖 3DMM 的传统网格解决方案通常以牺牲表征能力为代价，但渲染效率较高。另一方面，基于 NeRF 的方法提高了表征能力和保真度，却牺牲了效率。

Gaussian splatting methods attempt to resolve these trade-offs by offering both high representation power and efficiency with real-time rendering. However, their efficiency nonetheless depends on the number of rasterized Gaussians, and the time needed to optimize and prepare the head (the enrollment time) is often high.

高斯溅射（Gaussian splatting）方法试图通过提供高表征能力和实时渲染效率来解决这些权衡问题。然而，它们的效率仍然取决于光栅化高斯的数量，且优化和准备头部模型所需的时间（注册时间）通常很高。

In light of this, we propose EGGHead, which introduces a novel Gaussian allocation algorithm that efficiently assigns Gaussians to the head’s semantic regions given a fixed Gaussian budget. Our algorithm reduces the number of Gaussians needed while minimizing the impact on representation power.

鉴于此，我们提出了 EGGHead，它引入了一种新颖的高斯分配算法，能够在给定的高斯预算下，高效地将高斯分配给头部的语义区域。我们的算法在减少所需高斯数量的同时，最大限度地减小了对表征能力的影响。

EGGHead operates in a single forward pass and outperforms previous single-image face modeling methods in reconstruction quality, novel view synthesis, and rendering speeds.

EGGHead 通过单次前向传播运行，在重建质量、新视角合成和渲染速度方面均优于以往的单图像人脸建模方法。