Video summarization based on extracted key position of spotted objects
Zaur Fataliyev, David K. Han, Yadigar İmamverdiyev, Hanseok Ko · 2015
This paper proposes a novel fusion method for summarization of surveillance videos based on extracted key positions of spotted objects in observed area. Accumulated energy of object is calculated by analyzing its motion pattern for key position extraction. The method allows to summarize long videos into a single index frame. Experimental results demonstrate its effectiveness.