Scal3R：面向大规模三维重建的可扩展测试时训练
Scal3R: Scalable Test-Time Training for Large-Scale 3D Reconstruction
用一组轻量神经子网络作为可压缩的全局上下文记忆，并在测试时以自监督目标快速适配，为长视频前馈重建提供了一种不显著增加算力的长程记忆机制，可直接迁移到 SLAM、长序列位姿估计与流式三维重建。
原文依据 lightweight neural sub-networks that are rapidly adapted during test time via self-supervised objectives
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论本文研究从长视频序列中进行大规模三维场景重建的任务。近期的前馈重建模型通过直接从 RGB 图像回归三维几何，无需显式三维先验或几何约束，已展现出有前景的结果。然而，由于记忆容量有限且无法有效捕捉全局上下文线索，这些方法在长序列上往往难以保持重建精度与一致性。相比之下，人类能够自然地利用对场景的全局理解来指导局部感知。受此启发，我们提出一种新颖的神经全局上下文表示，可高效压缩并保留长程场景信息，使模型能够利用丰富的上下文线索来提升重建精度与一致性。该上下文表示由一组轻量神经子网络实现，这些子网络在测试阶段通过自监督目标被快速适配，从而在不引入显著计算开销的前提下大幅提升记忆容量。在包括 KITTI Odometry 与 Oxford Spires 数据集在内的多个大规模基准上的实验表明，我们的方法在处理超大规模场景方面有效，在保持效率的同时取得了领先的位姿精度与最优的三维重建精度。代码将发布于 https://zju3dv.github.io/scal3r。
英文原文摘要 1468 字符
This paper addresses the task of large-scale 3D scene reconstruction from long video sequences. Recent feed-forward reconstruction models have shown promising results by directly regressing 3D geometry from RGB images without explicit 3D priors or geometric constraints. However, these methods often struggle to maintain reconstruction accuracy and consistency over long sequences due to limited memory capacity and the inability to effectively capture global contextual cues. In contrast, humans can naturally exploit the global understanding of the scene to inform local perception. Motivated by this, we propose a novel neural global context representation that efficiently compresses and retains long-range scene information, enabling the model to leverage extensive contextual cues for enhanced reconstruction accuracy and consistency. The context representation is realized through a set of lightweight neural sub-networks that are rapidly adapted during test time via self-supervised objectives, which substantially increases memory capacity without incurring significant computational overhead. The experiments on multiple large-scale benchmarks, including the KITTI Odometry [??] and Oxford Spires [??] datasets, demonstrate the effectiveness of our approach in handling ultra-large scenes, achieving leading pose accuracy and state-of-the-art 3D reconstruction accuracy while maintaining efficiency. Code will be available at https://zju3dv.github.io/scal3r.