Overview

Along with increased demand for access to distributed information and communication via the Internet, wireless PDA devices and cellular phones, comes a tremendous need for information providers to adapt to various user demands, not only in terms of the diversity of content, but also in how that content is delivered.

At a high level the users may require that a given media set conform to certain specifications for language or may benefit from being presented data with a given level of abstraction. If, for example, a user is searching an international news database of videos, English subtitling may be required, while a thumbnail representation of key-frames may not be required, but ultimately benefit the user by speeding the browsing process. We will refer to the need for the system to respond to requirements defined by the user, including task level requirements as user-driven adaptation.

At a system level, there are a number of different factors which must be considered in the process of requesting and subsequently passing media objects (text, sound, images and video). In an environment where there are fixed levels network and system performance are guaranteed and the client hardware and software are known and do not change, we can design multimedia informations systems to easily deliver information as needed. However, with the growing number of network, hardware and software options we have to access information electronically, we find that a single solution for delivery is typically not sufficient. Networks have widely varying performance parameters which effect bandwidth and QOS (quality of service) and the user's client terminal has hardware and software configurations that have very specific functionality and can affect its ability to make use of different media representations. The ability of a system to respond appropriately to network constraints will be referred to as network-driven adaptation and its ability to respond to the client requirements will be referred to as client-driven adaptation.

By far, the majority of the literature which addresses these problems has dealt with the system level issues related to network- and client-driven adaptation, without particular regard for the information content of the data itself. In particular, research has focused on the requirements of low bandwidth channels and with reducing the amount of data to be transmitted over them by trading CPU cycles for bandwidth in order to reduce the amount of data to be sent. This can be done by providing lossy compression of the data or organizing the data such that only the essential data is transmitted initially and other data is transmitted on an as needed basis. This focus is justified by the fact that although the affordability of computing power has increased rapidly, the speed of various network connections such as wireless networks, phone lines and even dedicated lines has not kept with the growth in multimedia demand. Although various topics related to user-driven adaptation, such as media abstraction, analysis and understanding have been of interest for some time, they are far less well defined, and are much more closely tied to the demands of the user than parameters of the specific task or of the environment.

We define the scope of network and client adaptation problems, identify salient issues, highlight research directly and indirectly related to media adaptation and make recommendations for continued efforts. In particular, we will focus on what data specific capabilities can be provided which are both well defined and result in a measurable improvement for the end user.