Semi-automated Hyperlink markup for archived video

David Stotts, Jason McC. Smith
2002 Proceedings of the thirteenth ACM conference on Hypertext and hypermedia - HYPERTEXT '02  
One of the problems keeping video from being a fully first-class data component of hypermedia documents is the difficulty of treating the objects depicted in video as identifiable, linkable content. Rather, video tends to be manipulated as frames of pixels with no further subdivisions. When link markup is done on video streams, it is done manually frame-by-frame. We have been working with OvalTine, a system for tracking objects in video streams so that hypermedia link anchors can be associated
more » ... ith the objects in the video frames. In this work we show how the OvalTine tracking system can be used to do automated link markup of video streams. While our previous work with OvalTine presented object tracking in real-time streams, we concentrate here on markup of stored (archived) data. Our results will allow hypermedia structure to be generated and added to large digital libraries of video data.
doi:10.1145/513338.513367 dblp:conf/ht/StottsS02 fatcat:fe5xlynmk5aedc3kblmtlbrbry