AI RESEARCH

TrajLoom: Dense Future Trajectory Generation from Video

arXiv CS.CV

ArXi:2603.22606v1 Announce Type: new Predicting future motion is crucial in video understanding and controllable video generation. Dense point trajectories are a compact, expressive motion representation, but modeling their future evolution from observed video remains challenging. We propose a framework that predicts future trajectories and visibility from past trajectories and video context.