解决视频流3D重建中的几何漂移问题，提升稳定性。
Stabilizing Streaming Video Geometry via Dynamic Feature Normalization
- 动态归一化模块实时调节特征统计量，防止深度尺度漂移。
- 仅增加2%参数，实现比现有方法高14%的时序稳定性。
- 适合自动驾驶、机器人等需要连续3D感知的应用场景。
从连续的RGB视频流中稳定估计3D几何结构对自动驾驶、具身智能和大规模重建等实际应用至关重要。尽管现代单目几何基础模型在单帧上表现优异，但在连续输入下存在严重时序不一致，主要表现为尺度-偏移漂移。通过针对性的实证分析，我们发现其根源在于潜在特征统计量（均值与方差）的波动，直接影响预测深度的尺度与偏移。基于此洞察，我们提出动态特征归一化（DyFN），一种轻量级、因果递归模块，可动态且稳健地调节特征统计量，从而维持长时间的几何稳定性。通过仅微调DyFN（额外2%参数），同时冻结主干网络，即可适配强大的预训练单目几何模型于视频流任务，实现时序一致性而不损失单帧精度。在四个基准上的大量实验表明，DyFN有效消除断裂分层和位置抖动等时序伪影，达到当前最佳时序稳定性，较先前流式方法最高提升14%，甚至优于更重的非因果视频基线。
原文摘要 · Abstract (English)
Consistent 3D geometry estimation from streaming RGB input is crucial for real-world applications such as autonomous driving, embodied AI, and large-scale reconstruction. While modern monocular geometry foundation models achieve strong single-image accuracy, they exhibit severe temporal inconsistency on continuous input, notably dominated by scale--shift drifting. Through targeted empirical analysis, we trace this instability to its root cause: fluctuations in latent feature statistics, whose mean and variance directly determine the predicted depth's scale and shift. Building on this insight, we introduce Dynamic Feature Normalization (DyFN), a lightweight, causal recurrent module that dynamically and robustly modulates feature statistics to maintain stable geometry over time. We adapt powerful pretrained monocular geometry models for streaming by finetuning only DyFN, a mere 2\% additional parameters, while keeping the backbone frozen, thereby achieving temporal consistency without compromising single-image accuracy. Extensive experiments across four benchmarks show that DyFN effectively eliminates temporal artifacts such as disjointed layering and positional jitter, and achieves state-of-the-art temporal stability, improving over prior streaming methods by up to 14\% and even outperforming heavier non-causal video baselines. Project Page: https://shawlyu.github.io/DyFN
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。