Hierarchical brushing in a collection of video data

The amount of digital video and associated metadata being generated and stored is increasing rapidly. Given the complex spatial and temporal structure of video information it is a formidable challenge to provide compact and human-readable representations of such content. The solution has to be versatile in order to satisfy different user needs, such as browsing, zooming (looking for something specific), discovering or recording access patterns performing partial searches that can be re-used later, etc. We present a new approach that addresses many of the difficulties. Our representation provides at-a-glance high-level overview of the video collection and also serves as a navigational tool in that collection. While the user navigates to finer-level representations in the video collection our visualization allows him to maintain a sense of location and context within the collection. We introduce the movieDNA as an abstraction to visualize interesting features in a video. Our compact, yet flexible, representation is applicable to any type of linear data. In order to access even larger amounts of data in one view and to express several levels of granularity and compactness we extend the movieDNA to a hierarchical movieDNA. Hierarchical brushing enables browsing, navigating and visualizing several semantic levels of video content. We discuss hierarchical brushing of video content for various user scenarios and present examples demonstrating the versatility of our approach.