Nowadays the information society becomes more and more used to deal with multimedia data in various domains of life. Furthermore, multimedia online services represent one of the most significant markets of information and communication technologies. This leads to a need for concepts supporting not only the interaction with static media such as text, images or graphics, but also with dynamic media such as video or audio. In the recent years many multimedia applications have been developed providing video, but they mostly offer restricted capabilities of a video camera recorder to interact with it. However, the naive user would like to use the dynamic media with the same extent as the traditional static media. Applications such as video on demand, interactive television or multimedia information systems will benefit from technologies providing such new interaction facilities.

The introduction of sensitive regions, that are an extension of the anchor concept common in hypermedia models, offers new ways to interact with the medium video. Sensitive regions are used to define regions of interest in the video and to attach application-dependent semantics to the objects delineated within those regions. This allows to retrieve additional semantic information of the objects contained in a video.

The application MOVieGoer considers the hypermedia aspect of sensitive regions - the use as a dynamic anchor for retrieval from video. It was established in a joint co-operation with CCLRC Rutherford Appleton Laboratory, UK. The concept of MOVieGoer as a WWW-application provides the following features:

• An extended anchor concept allows to attach a hyperlink to a video object.
• Usage of sensitive regions to define hot spots within a video.
• Enables users to interact with objects occurring in the video. This allows the invocation of further actions attached to the selected object or to retrieve additional information about it. Users of MOVieGoer are for example able to click on an actor in a video and the WWW browser will respond by displaying information appropriate to that actor in that context.
• Based on standardized technology and can be applied within various application scenarios.
• Allows the definition of links without changing the video format or referenced WWW information.

• Display of sensitive regions on the client side (presentation) to inform the user about selectable objects within the video depends on the application context, but is mostly desirable.
• How long a video object is visible should be considered by an author. The time between the appearance and the disappearance should not be too short as otherwise the user is unable to select an object immediately.
• The majority of users prefer the video being stopped after a click occurred. It is an additional information to confirm the click of the user and they would miss something as several shots went by while viewing additional information presented by the WWW-browser.

To summarize, the concept of MOVieGoer enables to embed hyperlinks within dynamic media, i.e. video. It supports the content based integration of digital TV and WWW and is a further step within the convergence of television and computer systems into a single, interactive, digital multimedia information infrastructure.