Large-scale dynamic relational data\nvisualization has attracted considerable research\nattention recently. We introduce dynamic data\nvisualization into the multimedia domain, and present\nan interactive and scalable system, VideoMap, for\nexploring large-scale video content. A long video or\nmovie has much content; the associations between the\ncontent are complicated. VideoMap uses new visual\nrepresentations to extract meaningful information from\nvideo content. Map-based visualization naturally and\neasily summarizes and reveals important features and\nevents in video. Multi-scale descriptions are used\nto describe the layout and distribution of temporal\ninformation, spatial information, and associations\nbetween video content. Firstly, semantic associations\nare used in which map elements correspond to video\ncontents. Secondly, video contents are visualized\nhierarchically from a large scale to a fine-detailed\nscale. VideoMap uses a small set of sketch gestures to\ninvoke analysis, and automatically completes charts by\nsynthesizing visual representations from the map and\nbinding them to the underlying data. Furthermore,\nVideoMap allows users to use gestures to move and\nresize the view, as when using a map, facilitating\ninteractive exploration. Our experimental evaluation of\nVideoMap demonstrates how the system can assist in\nexploring video content as well as significantly reducing\nbrowsing time when trying to understand and find\nevents of interest.
Loading....