Long-range trajectories from global and local motion representations

• A system that extracts long-range trajectories from different scenarios is proposed.
• A rich fine-to-coarse representation of flow vectors is obtained.
• A new data-driven flow vector outlier removal technique is presented and justified.
• A global re-correlation algorithm at the tracklet-level is formulated.
• Our system largely outperforms state-of-the-art methods for motion segmentation.

Motion is a fundamental cue for scene analysis and human activity understanding in videos. It can be encoded in trajectories for tracking objects and for action recognition, or in form of flow to address behavior analysis in crowded scenes. Each approach can only be applied on limited scenarios. We propose a motion-based system that represents the spatial and temporal features of the flow in terms of long-range trajectories. The novelty resides on the system formulation, its generic approach to handle scene variability and motion variations, motion integration from local and global representations, and the resulting long-range trajectories that overcome trajectory-based approach problems. We report the results and conclusions that state its pertinence on different scenarios, comparing and correlating the extracted trajectories of individual pedestrians, manually annotated. We also propose an evaluation framework and stress the diverse system characteristics that can be used for human activity tasks, namely on motion segmentation.

Figure optionsDownload high-quality image (170 K)Download as PowerPoint slide