Combining multiple video frames and camera views significantly improves luggage re-identification in cluttered real-world environments, outperforming single-image matching approaches used in prior work.
This paper tackles luggage re-identification across multiple camera views in real-world settings like airports and train stations.