ELITE：通过学习式初始化与测试时生成自适应从单目视频高效构建高斯头部化身
ELITE: Efficient Gaussian Head Avatar from a Monocular Video via Learned Initialization and Test-time Generative Adaptation
用前馈 Mesh2Gaussian 先验模型快速初始化高斯头部化身，再用渲染引导的单步扩散增强器在测试时补全细节，为单目视频化身合成提供了「3D 先验初始化 + 2D 生成微调」的高效范式，可直接迁移到其他可动画高斯化身或人体重建任务。
原文依据 a rendering-guided single-step diffusion enhancer that restores missing visual details, grounded on Gaussian avatar renderings
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论我们提出 ELITE，一种通过学习式初始化与测试时生成自适应从单目视频高效合成高斯头部化身的方法。现有工作要么依赖 3D 数据先验，要么依赖 2D 生成先验，以补偿单目视频中缺失的视觉线索。然而，3D 数据先验方法往往难以在真实场景中泛化，而 2D 生成先验方法计算开销大且容易产生身份幻觉。我们发现这两种先验之间存在互补的协同效应，并据此设计了一个高效系统，在保持强真实场景泛化能力的同时实现高保真、可动画的化身合成。具体而言，我们引入前馈的 Mesh2Gaussian 先验模型（MGPM），实现高斯化身的快速初始化。为进一步弥合测试时的域间差距，我们设计了测试时生成自适应阶段，同时利用真实图像与合成图像作为监督。与以往缓慢且易产生幻觉的完整扩散去噪策略不同，我们提出渲染引导的单步扩散增强器，以高斯化身渲染结果为基准恢复缺失的视觉细节。实验表明，ELITE 生成的化身在视觉上优于现有工作，即使在富有挑战性的表情下也是如此，同时合成速度比 2D 生成先验方法快 60 倍。
英文原文摘要 1349 字符
We introduce ELITE, an Efficient Gaussian head avatar synthesis from a monocular video via Learned Initialization and TEst-time generative adaptation. Prior works rely either on a 3D data prior or a 2D generative prior to compensate for missing visual cues in monocular videos. However, 3D data prior methods often struggle to generalize in-the-wild, while 2D generative prior methods are computationally heavy and prone to identity hallucination. We identify a complementary synergy between these two priors and design an efficient system that achieves high-fidelity animatable avatar synthesis with strong in-the-wild generalization. Specifically, we introduce a feed-forward Mesh2Gaussian Prior Model (MGPM) that enables fast initialization of a Gaussian avatar. To further bridge the domain gap at test time, we design a test-time generative adaptation stage, leveraging both real and synthetic images as supervision. Unlike previous full diffusion denoising strategies that are slow and hallucination-prone, we propose a rendering-guided single-step diffusion enhancer that restores missing visual details, grounded on Gaussian avatar renderings. Our experiments demonstrate that ELITE produces visually superior avatars to prior works, even for challenging expressions, while achieving 60x faster synthesis than the 2D generative prior method.