SR3R：重新思考基于前馈高斯泼溅的超分辨率三维重建
SR3R: Rethinking Super-Resolution 3D Reconstruction With Feed-Forward Gaussian Splatting
将三维超分辨率从逐场景优化转为前馈映射，利用大规模多场景数据学习三维高频先验，可直接与任意前馈3DGS骨干网络结合，提升稀疏低分辨率输入下的重建保真度与跨场景泛化能力。
原文依据 We propose to reformulate 3DSR as a direct feed-forward mapping from sparse LR views to HR 3DGS representations
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论三维超分辨率（3DSR）旨在从低分辨率（LR）多视角图像重建高分辨率（HR）三维场景。现有方法依赖密集的低分辨率输入和逐场景优化，这限制了用于构建高分辨率三维高斯泼溅（3DGS）的高频先验只能继承自预训练的二维超分辨率（2DSR）模型。这严重制约了重建保真度、跨场景泛化和实时可用性。我们提出将三维超分辨率重新表述为从稀疏低分辨率视图到高分辨率三维高斯泼溅表示的直接前馈映射，使模型能够从大规模多场景数据中自主地学习三维特定的高频几何与外观。这从根本上改变了三维超分辨率获取高频知识的方式，并使其能够稳健地泛化到未见场景。具体而言，我们引入了SR3R，一个前馈框架，通过所学习的映射网络直接从稀疏低分辨率视图预测高分辨率三维高斯泼溅表示。为进一步提升重建保真度，我们引入了高斯偏移学习和特征细化，以稳定重建并锐化高频细节。SR3R即插即用，可与任何前馈三维高斯泼溅重建骨干网络配对：骨干网络提供低分辨率三维高斯泼溅支架，SR3R将其上采样为高分辨率三维高斯泼溅。在三个三维基准上的大量实验表明，SR3R超越了最优的三维超分辨率方法，并实现了强大的零样本泛化，甚至在未见场景上优于最优的逐场景优化方法。
英文原文摘要 1553 字符
3D super-resolution (3DSR) aims to reconstruct high-resolution (HR) 3D scenes from low-resolution (LR) multi-view images. Existing methods rely on dense LR inputs and per-scene optimization, which restricts the high-frequency priors for constructing HR 3D Gaussian Splatting (3DGS) to those inherited from pretrained 2D super-resolution (2DSR) models. This severely limits reconstruction fidelity, cross-scene generalization, and real-time usability. We propose to reformulate 3DSR as a direct feed-forward mapping from sparse LR views to HR 3DGS representations, enabling the model to autonomously learn 3D-specific high-frequency geometry and appearance from large-scale, multi-scene data. This fundamentally changes how 3DSR acquires high-frequency knowledge and enables robust generalization to unseen scenes. Specifically, we introduce SR3R, a feed-forward framework that directly predicts HR 3DGS representations from sparse LR views via the learned mapping network. To further enhance reconstruction fidelity, we introduce Gaussian offset learning and feature refinement, which stabilize reconstruction and sharpen high-frequency details. SR3R is plug-and-play and can be paired with any feed-forward 3DGS reconstruction backbone: the backbone provides an LR 3DGS scaffold, and SR3R upscales it to an HR 3DGS. Extensive experiments across three 3D benchmarks demonstrate that SR3R surpasses state-of-the-art (SOTA) 3DSR methods and achieves strong zero-shot generalization, even outperforming SOTA per-scene optimization methods on unseen scenes.