AUTHOR(S): 

Klimis Ntalianis, Nikos Mastorakis

TITLE

Increasing the Accuracy of Video Abstraction from Multiple Sources in the Internet of Things

KEYWORDS

Video Abstracts, Internet of Things, Interpolation, Accuracy Increase

ABSTRACT

In this paper a multiple sources video abstraction scheme is proposed for summarizing video content in the Internet of Things IoT). The proposed scheme borrows concepts from the IoT in case of events’ recording from different cameras around a specific geo-location. In particular the proposed abstraction method assumes that frame trajectories are created and an interpolation scheme is designed to optimally approximate these trajectories. The selected points on every trajectory are extracted as key-frames. Key-frames from all different sources are put in chronological order to produce the video abstract. In case a user demands more accuracy, he/she can increase it by refining the number of key-frames. In this case an optimal accuracy increasing approach is proposed and the video abstract is tuned according to the needs of the user. Experimental results on real world videos exhibit the promising performance of the proposed scheme.

Cite this paper

Klimis Ntalianis, Nikos Mastorakis. (2017) Increasing the Accuracy of Video Abstraction from Multiple Sources in the Internet of Things. International Journal of Internet of Things and Web Services, 2, 89-95