Flow3r：面向可扩展视觉几何学习的因子化光流预测
Flow3r: Factored Flow Prediction for Scalable Visual Geometry Learning
把稠密 2D 对应（flow）当作无标签监督信号，并用「几何 latent × 位姿 latent」的因子化光流预测来同时驱动场景几何与相机运动的学习，可直接迁移到现有前馈式 3D/4D 重建架构上，用于在无标注单目视频上扩展训练。
原文依据 predicting flow between two images using geometry latents from one and pose latents from the other
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论当前的前馈式 3D/4D 重建系统依赖稠密几何与位姿监督，这类监督在大规模场景下获取成本高昂，且在动态真实场景中尤为稀缺。本文提出 Flow3r，一个以稠密 2D 对应（「光流」）作为监督来增强视觉几何学习的框架，使其能够从无标注单目视频中进行可扩展训练。我们的关键洞见在于，光流预测模块应当被因子化：利用来自其中一张图像的几何隐变量与来自另一张图像的位姿隐变量来预测两图之间的光流。这种因子化直接引导场景几何与相机运动的学习，并自然地扩展到动态场景。在受控实验中，我们表明因子化光流预测优于其他设计方案，且性能随无标注数据规模一致提升。将因子化光流集成到现有视觉几何架构中，并使用 80 万段无标注视频进行训练，Flow3r 在涵盖静态与动态场景的八个基准上取得最优结果，其中在标注数据最为稀缺的真实动态视频上增益最大。
英文原文摘要 1125 字符
Current feed-forward 3D/4D reconstruction systems rely on dense geometry and pose supervision - expensive to obtain at scale and particularly scarce for dynamic real-world scenes. We present Flow3r, a framework that augments visual geometry learning with dense 2D correspondences ('flow') as supervision, enabling scalable training from unlabeled monocular videos. Our key insight is that the flow prediction module should be factored: predicting flow between two images using geometry latents from one and pose latents from the other. This factorization directly guides the learning of both scene geometry and camera motion, and naturally extends to dynamic scenes. In controlled experiments, we show that factored flow prediction outperforms alternative designs and that performance scales consistently with unlabeled data. Integrating factored flow into existing visual geometry architectures and training with 800K unlabeled videos, Flow3r achieves state-of-the-art results across eight benchmarks spanning static and dynamic scenes, with its largest gains on in-the-wild dynamic videos where labeled data is most scarce.