用多参考帧定位解决长视频3D重建漂移问题，效果显著提升。
Scal3R: Learning Efficient Multi-Relative Pose Query for Scalable Online 3D Reconstruction
- 通过多个历史关键帧查询相对位姿，替代固定首帧锚点。
- 在KITTI上平均ATE降低60%以上，8小时单卡完成训练。
- 适合长时序3D重建任务，尤其对实时系统有实用价值。
在线3D重建模型在长视频上表现差，因相对固定首帧的位姿回归导致超出训练分布的外推，微小漂移累积引发几何崩溃。我们观察到每帧深度保持稳定，骨干网络局部几何完好，仅全局位姿头失效。基于此解耦现象，提出Scal3R：将在线重建重构为多参考相对位姿查询。引入约1%参数量的轻量可学习标记，通过非对称注意力注入完全冻结的骨干网络，实现相对于多个历史关键帧的位姿查询。结合具备回环闭合功能的在线位姿图优化系统，抑制远距离漂移。Scal3R在单个GPU上8小时内收敛，在KITTI上平均绝对轨迹误差（ATE）相比在线基线降低超过60%，并在Virtual KITTI、Sintel、TUM-Dynamic、ScanNet和7-Scenes上达到当前最佳性能。
原文摘要 · Abstract (English)
Online 3D reconstruction models perform poorly on long videos. This happens because regressing poses relative to a fixed first-frame anchor forces extrapolation far beyond the training distribution. Small drifts accumulate and amplify into significant geometric collapse. However, we observe that per-frame depth remains stable throughout this failure. The backbone's local geometry remains intact; only the global pose head breaks down. Motivated by this decoupling, we introduce Scal3R. This approach reformulates online reconstruction as multi-reference relative pose querying. We use lightweight learnable tokens, which make up about ~1% of the parameters, and inject them into a completely frozen backbone via asymmetric attention. This setup queries poses relative to multiple past keyframes. An online pose-graph optimization system with loop closure suppresses long-range drift. Scal3R reaches convergence in 8 hours on a single GPU. It reduces the average ATE by over 60% on KITTI compared to the online baseline. It also achieves state-of-the-art performance across Virtual KITTI, Sintel, TUM-Dynamic, ScanNet, and 7-Scenes. Project page: https://linjohnss.github.io/scal3r/
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。