Robust Subband / Wavelet Coding and Transmission of Images and Video

In this text we investigate robust subband/wavelet coding and transmission of images and video. The text is divided into two parts. In the first part we focus on robust encoding of image and video sources by domain partitioning. The goal is to optimally divide the signal domain and encode the different components independently so that missing components can be easily estimated from the received ones upon transmission. We show that such a scheme can provide relatively good results even when simple estimation methods are used. Further performance improvement can be achieved through unequal loss protection provided by channel coding or intelligent congestion control. In the second part we focus on robust streaming of scalable video. Our approach is based on the recently developed MC-EZBC coder. We use channel coding to create multiple descriptions from MC-EZBC video bit streams for robust streaming. We also propose a motion compensated estimation method to recover parts of the video affected by high loss. Finally, we describe the MC-EZBC video streaming experiments performed on the MPEG-21 multimedia test bed.