A 3D-DCT and Convolutional FEC Approach to Agile Video Streaming

On today's Internet, including both wide area networks and wireless access networks, multimedia content (images, audio and video) represents a dominating fraction of network data traffic. Efficient encoding and transmission of multimedia data are of considerable importance to smooth operation of the network as well as quality of experience of network users. Wireless users particularly prefer algorithms that are both robust, and lightweight in terms of computation and communication. This work takes a holistic approach to source compression and network transmission of videos. A 3D-DCT module is employed for spatio-temporal compression of pixel cubes, working in concert with a convolutional network coding module, for unequal error protection and forward error correction. We outline the overall framework for such a joint algorithm design, discuss detailed algorithm blocks, while many interesting design choices are still open for future exploration.