Quality evaluation of 3D video for QoE management in media networks

Abstract

The emergence of 3D video technologies and their foreseeable applications in Internet Protocol Television (IPTV) and video delivery systems using encoded 3D video contents over packet networks raises the question of how to manage the Quality of Experience (QoE) across diverse lossy channels. Of particular importance for QoE management is the problem of measuring the impact of data losses in packetized 3D video information and how it affects the quality experienced by end users, when the content is rendered and presented at their premises. Previous work by the authors showed that it is possible to model the perceived quality degradations through the use of artificial neural networks receiving as inputs several parameters describing the packet loss events. The first stage of such model was specifically developed for packetized 3D video in texture-plus-depth format where only the depth information was prone to transmission errors. This article presents an extension of the previous model by including the effect of texture information losses along with other specific aspects associated with the dual nature of this type of data loss. The validity of the model is verified through the use of extensive simulations and comparisons between real and estimated values of a recently proposed 3D video quality measure.