LongStream: 长序列流式自回归视觉几何
LongStream: Long-Sequence Streaming Autoregressive Visual Geometry
AI总结：
LongStream通过抛弃首帧锚点、引入正交尺度学习和缓存一致性训练，解决长序列流式3D重建中的姿态漂移和注意力偏差问题，实现千帧级稳定重建。
摘要：
长序列流式3D重建仍是一个重大开放挑战。现有自回归模型在处理长序列时往往失效，因为它们将姿态锚定在第一帧，导致注意力衰减、尺度漂移和外推误差。我们引入LongStream，一种新型的尺度解耦流式视觉几何模型，用于在严格在线、未来不可见的设置下，实现数千帧的度量尺度场景重建。我们的方法三方面：首先，抛弃首帧锚点，预测关键帧相对姿态，将长程外推转化为常难度的局部任务。其次，引入正交尺度学习，完全解耦几何与尺度估计以抑制漂移。最后，我们识别Transformer中的注意力偏差问题，包括注意力沉底依赖和长期KV缓存饱和。我们提出缓存一致性训练结合定期缓存刷新。该方法抑制了超长序列中的注意力偏差和污染，减少了训练与推理之间的差距。实验表明，LongStream实现了最先进的性能，能够在18 FPS下实现千米级序列的稳定度量尺度重建。项目页面：https://3dagentworld.github.io/longstream/
英文摘要：
Long-sequence streaming 3D reconstruction remains a significant open challenge. Existing autoregressive models often fail when processing long sequences because they anchor poses to the first frame, leading to attention decay, scale drift, and extrapolation errors. We introduce LongStream, a novel gauge-decoupled streaming visual geometry model for metric-scale scene reconstruction across thousands of frames under a strictly online, future-invisible setting. Our approach is threefold. First, we discard the first-frame anchor and predict keyframe-relative poses. This reformulates long-range extrapolation into a constant-difficulty local task. Second, we introduce orthogonal scale learning. This method fully disentangles geometry from scale estimation to suppress drift. Finally, we identify attention bias issues in Transformers, including attention-sink reliance and long-term KV-cache saturation. We propose cache-consistent training combined with periodic cache refresh. This approach suppresses attention biases and contamination over ultra-long sequences and reduces the gap between training and inference. Experiments show that LongStream achieves state-of-the-art performance, enabling stable, metric-scale reconstruction over kilometer-scale sequences at 18 FPS. Project Page: https://3dagentworld.github.io/longstream/
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
