On quality evaluation of 3D video using Colour-plus-Depth and MDC

This paper presents a quality evaluation of three-dimensional Colour-plus-Depth videos over error prone IP networks and illustrates an overall cloud architecture in which such an evaluation platform may reside. For achieving video transmission both scalable and multiple-description coding techniques are utilized. On evaluating the impact of packet loss to the perceived 3D video quality, asymmetric packet loss was used for characterizing path diversity. Mean Opinion Score (MOS) for the multiple description scheme is affected by both the encoding parameters and the sensed packet loss rate at each 3D description. The effect of 3D video reconstruction through interpolation using a single yet reliably delivered description was compared against using the overall number of delivered descriptions. Throughout this experiment it was found that almost all subjects tend to rate the first group of reconstructed 3D videos with higher scores when the network delivery mechanism undergoes severe packet loss rate conditions.