让多个物体按指定轨迹运动，精准定位不混淆。
TrajLoc: Trajectory-Attention Localization for Multi-Object Motion Control
- 用高斯热图替代交叉注意力，每个物体独立控制位置
- 20个物体同时控制时，轨迹误差降低51%，清晰度提升4.3分贝
- 适用于真实场景，兼容不同视频生成模型
在图像到视频生成中，控制多个物体的运动需保持物体身份并遵循各自轨迹，尤其在物体数量多、路径交叉或遮挡时更具挑战。现有方法将多条轨迹混入共享密集条件信号，难以在密集场景中维持物体对应关系。本文提出TrajLoc，直接在注意力层中为每个物体施加独立的空间约束：将每个物体令牌的交叉注意力权重替换为以目标位置为中心的高斯热图。同一物体令牌同时携带轨迹与深度信息，并通过首帧外观编码保持身份识别。在六个数据集上评估，最多支持20个物体同时控制，涵盖分布外真实场景，结果表明本方法在视觉保真度和轨迹遵循性上均显著优于最强基线。应用于两种不同架构（CogVideoX 5B 和 WaN 2.1 14B），平均提升4.3 dB PSNR，轨迹终点误差减少51%。
原文摘要 · Abstract (English)
Controlling the motion of multiple objects in image-to-video (I2V) generation requires preserving object identities while enforcing adherence to distinct target trajectories. This becomes particularly challenging as the number of objects increases and their paths intersect or occlude one another. Existing approaches entangle multiple trajectories within a shared, dense conditioning signal, making object-level correspondence difficult to preserve in crowded scenes. We depart from this paradigm and enforce a strict, per object spatial constraint that isolates instances independently. Our method, TrajLoc, achieves this directly within the attention layers by substituting the cross-attention weights of each object token with a Gaussian heatmap centered on its target location at every frame. The same per object token interface carries trajectory and depth through a learned embedding and preserves identity by encoding first frame appearance in place of an object token. Evaluations across six datasets, featuring up to 20 simultaneously controlled objects and out of distribution real world scenes, demonstrate that our method consistently improves both visual fidelity and trajectory adherence. Applied to two architecturally distinct backbones (CogVideoX 5B and WaN 2.1 14B), our approach achieves average gains of +4.3 dB PSNR and a 51% reduction in trajectory end point error compared to the strongest baselines. Project page: https://sela-omer.github.io/traj-loc/
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。