Distortion optimization in enriched video traces for End-to-End video quality enhancement

Video compression and streaming over lossy wireless networks is the current trend in telecommunication and encouraged by major carriers to increase their services portfolio. However, the technological challenges are still tackled to enable the delivery of high quality video. In our previous work, we developed a per stream distortion model that is based on the extended Gilbert model. We expanded our definition of the verbose video trace file with embedded coefficients derived from our distortion model. These coefficients are video quality descriptors that we associate with Quality of Service parameters that are measured in real-simulation time. In this work, we devise an optimization algorithm that is based on the distortion model to enhance the received video quality in End-to-End (E2E) communications. Our optimization algorithm is tested in multiple simulations to verify its validity. We report the results from our simulations that indicate significant improvements when this optimization algorithm is deployed in Edge Routers in conjunction with our video stream distortion coefficients and network metrics.