Abstract: An automated video editing and indexing system is taught which has application for such tasks as creating a presentation for a video magazine and processing collections of video material in general. The system extracts text either from closed captions (CC) contained in the analog video presentation, or extracted from the voice contained in the presentation by voice-to-text techniques, time-stamps the CC text in text files according to position in the video presentation, and digitizes the video presentation. The text files and digitized video are sent to an editing station, where the CC text is analyzed using Natural Language Processing and other techniques to determine topic changes in the presentation. Keyframes are selected to represent the topic changes, and become thumbnails useful in indexing; indexing meaning marking the video material at points of topic changes, and in some cases jumping the video presentation to the positions represented by the thumbnails.

Abstract: A tracking software module for tracking a moving image entity in a video display has an interface for receiving and processing a video data stream while the video presentation provided by the data stream is presented on a display as a series of bitmapped images having a moving image entity to be tracked. There is in the module a graphical user interface usable by an editor for creating, selecting and centering a predefined matrix of signature pixel positions on the moving entity to be tracked in a first frame of the video presentation, and for initiating a tracking process. A determination function determines the color values of the signature pixels in the first frame and stores the values as an image signature.

Abstract: A system and method for hypervideo object tracking in a compressed digital video environment includes a hypervideo authoring system and a hypervideo browser. The hypervideo authoring system permits one or more moving objects in an MPEG compressed video clip to be identified, tracked, refined, edited, and associated with a specified action with a minimum of user intervention. The hypervideo browser allows the MPEG video to be viewed and objects of interest to be selected by a user, thereby hyperlinking to additional information of interest.

Abstract: Streaming digital hypervideo including copious embedded hyperlinks is distributed upon a digital communications network from a hypervideo server, normally an Internet Service Provider, to multitudinous client subscribers/users/viewers (client SUVs). Some or all of the client SUVs receive the same hyperlinks at the same place in the streaming hypervideo. Some small fraction of the client SUVs selectively volitionally exercise a fraction of the total hyperlinks, causing an access in the background of the unfolding hypervideo across the digital communications network to yet another server commonly called a "Video On Web server", or "VOW server". The VOW sever interprets each hyperlink request in consideration of (i) the identity of the exercising client SUV and, most commonly, (ii) additional data of a demographic, socioeconomic, credit, viewing preference, security and/or past hyperlinking history nature.