2016 IEEE International Conference on Multimedia & Expo Workshops (ICMEW) (2016)
Seattle, WA, USA
July 11, 2016 to July 15, 2016
Ahmed Aldahdooh , LUNAM Universite de Nantes, IRCCyN UMR CNRS 6597, France
Marcus Barkowsky , LUNAM Universite de Nantes, IRCCyN UMR CNRS 6597, France
Patrick Le Callet , LUNAM Universite de Nantes, IRCCyN UMR CNRS 6597, France
Error resilience (ER) and error concealment (EC) are important tools in video transmission over lossy channels. Notably, when maximizing Quality of Experience (QoE) instead of Quality of Service (QoS), some designs of multiple description coding (MDC) schemes are promising. MDC is categorized into different types and, in this paper, we focus on temporal MDC techniques. In video coding, such as MDC, the quality of received video depends on the content type and on the error concealment technique. Transmitting the video sequences using either single description coding (SDC) or MDC (2-MDC/4-MDC) poses a challenge in terms of gains and losses. In this paper, a framework that predicts either to use SDC or to use MDC to transmit the video sequences using underlying content characteristics is introduced. The ground truth of the prediction model is built using a subjective experiment's result. An SVM fit on 12 ultra high definition (UHD) sequences reaches a classification accuracy of 91% with only 3 features that are related to content features and the quality of service (QoS) features, i.e. the degradation effect of the error concealment technique of the coding scheme and it is measured using PSNR.
Bit rate, Encoding, Feature extraction, Quality of service, Video sequences, Predictive models, Resilience
A. Aldahdooh, M. Barkowsky and P. Le Callet, "Content-aware adaptive multiple description coding scheme," 2016 IEEE International Conference on Multimedia & Expo Workshops (ICMEW), Seattle, WA, USA, 2016, pp. 1-6.