DocumentCode
2101577
Title
A shortest path representation for video summarisation
Author
Porter, S.V. ; Mirmehdi, M. ; Thomas, B.T.
Author_Institution
Dept. of Comput. Sci., Bristol Univ., UK
fYear
2003
fDate
17-19 Sept. 2003
Firstpage
460
Lastpage
465
Abstract
A novel approach is presented to select multiple key frames within an isolated video shot where there is camera motion causing significant scene change. This is achieved by determining the dominant motion between frame pairs whose similarities are represented using a directed weighted graph. The shortest path in the graph, found using the A* search algorithm, designates the key frames. The overall method can be applied to extract a set of key frames which portray both the video content and camera motions, all of which are useful features for video indexing and retrieval.
Keywords
content-based retrieval; database indexing; directed graphs; feature extraction; video databases; A* search algorithm; camera motion; directed weighted graph; dominant motion; feature detection; frame pair similarities; isolated video shot; multiple key frames; shortest path representation; significant scene change; video content; video indexing; video retrieval; video summarisation; Algorithm design and analysis; Computer science; Content based retrieval; Data mining; Digital cameras; Indexing; Information retrieval; Layout; Multimedia systems; Videoconference;
fLanguage
English
Publisher
ieee
Conference_Titel
Image Analysis and Processing, 2003.Proceedings. 12th International Conference on
Print_ISBN
0-7695-1948-2
Type
conf
DOI
10.1109/ICIAP.2003.1234093
Filename
1234093
Link To Document