Ensuring Multimedia QoS Over Scarce Resource Networks

A serious bottleneck in multimedia E-learning is the non-availability of required bandwidth. While much research has been done on Multimedia compression and network resource optimization, multimedia delivery is still a big concern especially for scarce resource networks. The virtues of network aware video coding and application of domain specific compression techniques have not yet been fully explored. In this project we deal with the issue of intelligent transmission of important segments of the video sequence over the network. We incorporate the knowledge of the network conditions to determine how various parts of the video frames are encoded. An estimate of the available network bandwidth is obtained which is then distributed optimally between the different frame constituents based on their relative importance and motion by the Smart Decision Maker (SDM) module. We have devised a simple approach to classify the videos into visual objects and used Discrete Wavelet Transform based Color Embedded Zerotree Wavelet (CEZW) coding to obtain a scalable bitstream that provides dynamic response to changing network conditions. The above approach helps to reach an optimal trade off between perceptual quality of the video and the available network bandwidth. While this scheme is robust and operational for all low motion videos with identifiable visual objects, we take the example of educational videos to demonstrate the proposed scheme. The results are compared with the state of the art compression and transmission schemes over the network to illustrate the superior performance of our approach.