HumanBA：通过全局人体-相机解耦的人体感知光束法平差
HumanBA: Human-Aware Bundle Adjustment via Global Human-Camera Decoupling
将动态人体作为可用的路标点，通过运动解耦从关节轨迹中分离出相机诱导的伪静态分量，使其能安全地融入光束法平差，为单目视频中人体主导场景的相机位姿估计与全局人体重建提供了新思路。
原文依据 HumanBA subtracts the human-induced component from observed joint trajectories, isolating a camera-induced (pseudo-static) component that can be safely incorporated into bundle adjustment alongside background features.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论从单目视频中恢复全局人体与相机运动对于世界坐标系下的人体重建至关重要，但由于图像空间中运动相互纠缠而仍具挑战性。传统SLAM方法能够估计单目相机运动，但在以人体等前景物体为主的场景中会失效。一种常见的变通方法是掩蔽动态物体，然而当人体占据大部分视野或背景噪声过大时，该方法变得脆弱，导致跟踪不稳定并丢失约束。本文采取相反立场，将人体运动重新作为信息丰富的路标点纳入。我们提出HumanBA，一种人体感知的光束法平差框架，通过运动解耦将动态人体转化为可用的约束。HumanBA从观测到的关节轨迹中减去人体诱导分量，分离出可安全地与背景特征一同纳入光束法平差的相机诱导（伪静态）分量。为缓解全局人体估计中的噪声，HumanBA应用运动细化与运动感知可靠性加权。在EMDB和SLOPER4D基准上，我们展示了相机位姿估计的持续改进，并降低了全局人体重建误差，证明了将人体视为动态但信息丰富的路标点的优势。代码见 https://github.com/MartaYang/HumanBA。
英文原文摘要 1377 字符
Recovering global human and camera motion from monocular video is essential for world-coordinate human reconstruction but remains challenging due to entangled motions in image space. Traditional SLAM methods estimate monocular camera motion but fail in scenes dominated by foreground objects such as humans. A common workaround is to mask out dynamic objects, yet this approach becomes brittle when humans occupy most of the view or the background is too noisy, leading to unstable tracking and loss of constraints. This paper takes the opposite stance and reintegrates human motion as informative landmarks. We introduce HumanBA, a human-aware bundle adjustment framework that transforms dynamic humans into usable constraints via motion decoupling. HumanBA subtracts the human-induced component from observed joint trajectories, isolating a camera-induced (pseudo-static) component that can be safely incorporated into bundle adjustment alongside background features. To mitigate noise in global human estimates, HumanBA applies motion refinements and motion-aware reliability weighting. Across EMDB and SLOPER4D benchmarks, we show consistent improvements on camera pose estimation and reduce global human reconstruction error, demonstrating the benefits of treating humans as dynamic yet informative landmarks. Our code is available at https://github.com/MartaYang/HumanBA.