A computational approach to multimedia communication networks

Multimedia communication is becoming a common means of human interaction. This increasing amount of information transmitted over wireless and land-line communication channels has to be handled efficiently to avoid congestion and delays in the network service. In this work, a computational approach to multimedia data transmission is proposed. One of its basic assumptions is that in many situations the multimedia information could be decomposed into basic components or building blocks that represent a significant part of the data. To illustrate the capabilities of the proposed approach, a compression system for video streams is investigated and tested. The system is based on the observation that in most multimedia systems, events of recent history of the stream can be used to describe present and future details of the multimedia information. Experimental results indicate a high compression ratio of more than 100:1, obtaining almost lossless visual reproduction with less than 0.1% distortion. The proposed approach can be similarly used also for other components of the data stream. Our conclusion is that the proposed computational approach could be efficient in presently used communication systems.