任意分辨率任意几何：从多视图到多图像块
Any Resolution Any Geometry: From Multi-View To Multi-Patch
把高分辨率单张图切成多个图像块、用跨块注意力在共享骨干里做长程几何推理，是解决高分辨率几何估计中「局部细节 vs 全局一致」矛盾的通用手段；GridMix 训练期随机网格采样也可直接迁移到其他分块推理任务以提升块间一致性。
原文依据 Global coherence is enforced through cross-patch attention, which enables long-range geometric reasoning and seamless propagation of information across patches within a shared backbone.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论表面法线与深度的联合估计对于整体三维场景理解至关重要，然而由于在保留精细局部细节与维持全局一致性之间存在权衡，高分辨率预测仍然困难。为应对这一挑战，我们提出超分辨率几何 Transformer（URGT），它将视觉几何基础 Transformer（VGGT）改造为统一的多图像块 Transformer，用于单目高分辨率深度—法线估计。单张高分辨率图像被划分为若干图像块，这些图像块由预训练模型提供的粗略深度与法线先验进行增强，并在一次前向传播中联合处理以预测精细化的几何输出。全局一致性通过跨图像块注意力加以约束，该机制使得长程几何推理成为可能，并让信息在共享骨干内跨图像块无缝传播。为进一步增强空间鲁棒性，我们引入 GridMix 图像块采样策略，在训练过程中按概率采样网格配置，从而提升图像块间一致性与泛化能力。我们的方法在 UnrealStereo4K 上取得最优结果，同时改进深度与法线估计——将 AbsRel 从 0.0582 降至 0.0291，RMSE 从 2.17 降至 1.31，并将平均角度误差从 23.36 度降至 18.27 度——同时生成更锐利、更稳定的几何。所提出的多图像块框架还展现出强大的零样本与跨域泛化能力，并能有效扩展至极高分辨率，为高质量几何精细化提供了一种高效且可扩展的解决方案。
英文原文摘要 1618 字符
Joint estimation of surface normals and depth is essential for holistic 3D scene understanding, yet high-resolution prediction remains difficult due to the trade-off between preserving fine local detail and maintaining global consistency. To address this challenge, we propose the Ultra Resolution Geometry Transformer (URGT), which adapts the Visual Geometry Grounded Transformer (VGGT) into a unified multi-patch transformer for monocular high-resolution depth-normal estimation. A single high-resolution image is partitioned into patches that are augmented with coarse depth and normal priors from pre-trained models, and jointly processed in a single forward pass to predict refined geometric outputs. Global coherence is enforced through cross-patch attention, which enables long-range geometric reasoning and seamless propagation of information across patches within a shared backbone. To further enhance spatial robustness, we introduce a GridMix patch sampling strategy that probabilistically samples grid configurations during training, improving inter-patch consistency and generalization. Our method achieves state-of-the-art results on UnrealStereo4K, jointly improving depth and normal estimation---reducing AbsRel from 0.0582 to 0.0291, RMSE from 2.17 to 1.31, and lowering mean angular error from 23.36deg to 18.27deg--while producing sharper and more stable geometry. The proposed multi-patch framework also demonstrates strong zero-shot and cross-domain generalization and scales effectively to very high resolutions, offering an efficient and extensible solution for high-quality geometry refinement.