Querying multiple perspective video by camera metaphor

Suppose that, in a sports event, a user searches for and obtains video scenes that have "better" views of interesting objects from multiple-perspective video (a collection of mutually synchronized video data taken by a lot of cameras). Because members of the general public take videos and query them in such applications, one of the most important research issues is how to provide a mechanism by means of which a user can formulate appropriate queries easily while looking for and focusing on interesting objects. Another issue is how to search for video scenes that have "better" views of the objects. In this paper, we introduce a querying method based on the concept of "querying by camera", and we propose an algorithm that searches for video intervals that have spatio-temporally "better" views of the focused-on object.