[论文] WorldRecon: Generative Multi-View Mesh Reconstruction for Large-Scale ...

论文概要

研究领域: CV 作者: Mengchen Zhang, Yang Li, Chunshi Wang, Tong Wu, Chunchao Guo 发布时间: 2026-10-13 arXiv: 2610.18846

中文摘要

从无位姿多视图图像重建完整、高质量的大规模场景网格，既需要忠实几何锚定，又需要合理补全未观测或仅稀疏捕获的表面。我们提出 WorldRecon——一个生成式多视图重建框架，从无位姿 RGB 图像恢复完整的大规模 3D 场景网格。WorldRecon 使用重建模型从输入图像估计的粗略带噪全局点云和相机位姿作为全局脚手架，将其划分为重叠的空间块。为稳健利用这些不完美的几何先验，我们为原生 3D 生成模型 TRELLIS.2 配备了专门的基于重建的条件化架构，从局部点云裁剪和块相关图像观测生成每个块，产生稀疏结构、详细几何和 PBR 材质。为保持场景级一致性，我们引入邻居感知潜在修复和因果世界生成过程，在重叠块间传播结构、几何和材质上下文。这些设计共同将多视图生成式重建扩展到大规模场景，产生连贯统一的网格。大量实验证明 WorldRecon 取得了最先进的重建性能，在合成和真实室内场景上分别将 Chamfer 距离降低 14.9% 和 43.1%，F-score 分别提升 17.3% 和 12.0%。

原文摘要

Reconstructing complete, high-quality meshes of large-scale scenes from unposed multi-view images requires faithful geometric grounding while plausibly completing surfaces that are unobserved or only sparsely captured. We present WorldRecon, a generative multi-view reconstruction framework for recovering complete, large-scale 3D scene meshes from unposed RGB images. WorldRecon uses a coarse, noisy global point cloud and camera poses estimated from the input images by a reconstruction model as a global scaffold for whole-scene generation and partitions it into overlapping spatial blocks. To robustly leverage these imperfect geometric priors, we equip the native 3D generative model TRELLIS.2 with a dedicated reconstruction-based conditioning architecture that generates each block from its lo...

#论文 #arXiv #CV #小凯