Chapter Six: Egocentric video cosummarization
Abstract
Our objective in this chapter is to achieve egocentric video cosummarization. The problem is to summarize an egocentric (test) video at shot level with the help of some similar relevant (database) videos. Two different solutions are proposed. The first solution is based on bipartite graph matching and game theory. The visual similarity between a test video and a database video is obtained using a game-theoretic approach. The value of the expected payoff difference at mixed Nash equilibrium in a game between a database video shot and a test video shot (the two players) is deemed as the visual similarity score between them. A weighted bipartite graph is constructed across shots from a test, and ...
Get Graph Based Multimedia Analysis now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
