Abstract

In this paper, we develop a trajectory-based video retrieval system. To achieve it, we first design a spatio-temporal representation scheme for modeling the trajectories of moving objects in video databases. In addition, we present an efficient trajectory-based retrieval scheme using a k-warping distance algorithm. Finally, we propose a new signature-based access method which can provide good retrieval performance on a trajectory-based query. To show the usefulness of applying our above schemes to a real video database application, we implement a trajectory-based soccer video retrieval system because soccer video data generally contain a lot of motions according to a salient object called soccer ball. Our soccer video retrieval system provides graphic user interfaces for both indexing and retrieval. We also provide the performance analysis of our trajectory-based video retrieval scheme.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call