Video files are very complex objects. For many years, researchers developed models to allow for search-and-retrieval systems specific for these objects. Since the results of a query will be a set of videos or of segments of videos, their size may be prohibitive, and do not allow for pre-validation before downloading. Moreover, many features of the video files - for example the multiplicity of their formats - can make problems. We propose in this article an original approach based on the notions of metadata and video thumbnails, which can answer the problem of searching efficiently the video files, while allowing the quick transmission of information allowing to pre-view the results of the query before downloading them. Indeed, in our approach, the video thumbnails is integrated into the structure we propose to describe the video. This allows a powerful search within the various elements of the video. Our model is generic and flexible, so as to fit with the various structures of existing databases used for describing video files. © Springer-Verlag Berlin Heidelberg 2002.
CITATION STYLE
Vachon, E. (2002). A pre-viewing step in video retrieval. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 2490 LNCS, pp. 342–352). Springer Verlag. https://doi.org/10.1007/3-540-36128-6_19
Mendeley helps you to discover research relevant for your work.