Validation of a new full reference metric for quality assessment of mobile 3DTV content

3D video is expected to provide an enhanced user experience bringing more realism through depth impression. Therefore, quality assessment plays an important role when designing and optimizing 3D video processing methods and applications. In this paper, we propose and validate a novel quality metric based on binocular human visual system. Both visual distortions in the cyclopean view and perceptibility of depth are considered. Visual distortions are analysed in 3D-DCT domain taking into account the masking effects of the contrast sensitive function (CSF) and depth variability. The metric is especially tailored for mobile 3DTV applications as it takes into account the target display size and resolution and calculates disparity variations in local areas with relevant foveal size. The metric is designed by matching the mean opinion scores resulted from large-scale subjective tests with 3D videos with varying depth presence and amount of compression artefacts. It is further validated over another 3D video database with wider types of compression artefacts. The results show that the metric outperforms current metrics over different 3D video formats and compression methods.