[论文] SplatGuide: Geometric Priors from 3D Gaussians for Pose-Free Novel Vie...

论文概要

研究领域: CV 作者: Yejun Zhang, Zihan Wang, Xu Ji et al. (11 authors) 发布时间: 2026-08-17 arXiv: 2608.16863

中文摘要

从未姿态图像生成真实感新视角需要3D几何理解和合成未见过内容的能力。一个自然的策略是将前馈3DGS重建与多视图扩散结合。然而，先前管道最多从重建中提取一种信号（像素渲染或学习特征），没有利用每高斯可见性进行遮挡感知参考选择。这种信息断开使可渲染几何、可见性线索和学习特征未被利用。SplatGuide通过在不同角色中重用单个3DGS场景来弥合这一断开：渲染图像提供像素对齐的几何条件，每高斯源视图索引被渲染到目标视图投票图以进行遮挡感知参考选择，重建token通过交叉注意力提供特征级引导。三种信号都来自同一次重建前向传播。在RealEstate10K、DL3DV、Tanks-and-Temples和Mip-NeRF 360上，SplatGuide实现了SOTA的无姿态新视角合成。在RealEstate10K上，以中等数量的输入视图，它超越了真实姿态基线。

原文摘要

Generating photorealistic novel views from unposed images requires both 3D geometric understanding and the ability to synthesize unseen content. A natural strategy combines feed-forward 3DGS reconstruction with multi-view diffusion. Yet prior pipelines extract at most one signal from the reconstruction, either pixel rendering or learned features, while none exploits per-Gaussian visibility for occlusion-aware reference selection. This information disconnect leaves renderable geometry, visibility cues, and learned features unused. SplatGuide closes this disconnect by reusing a single 3DGS scene across three complementary roles. Rendered images provide pixel-aligned geometric conditioning. Per-Gaussian source-view indices are rendered into a target-view voting map for occlusion-aware referen...

--- *自动采集于 2026-08-19*

#论文 #arXiv #CV #小凯