A Statistical Approach to Packet Loss Concealment for Video

We have developed a statistical prediction technique to compensate for lost pixel blocks during real time transmission of video over packet-switched networks. The approach could as well be used for standard inter-frame coding. The method is based on the joint modeling of pixel statistics by Gaussian mixtures. Different EM variants for decreasing computational complexity and storage space are proposed. In the case of 4×4 luminance blocks, ourmethod increases performance by 2.3 dB when augmenting the number of mixture components from 1 to 64. We show that these results are statistically significant.©2009 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.Daniel Persson and Per Hedelin, A Statistical Approach to Packet Loss Concealment for Video, 2005, Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing, 2005. Proceedings. (ICASSP '05), 293-296