Stereoscopic video quality assessment based on stereo just-noticeable difference model

In this paper, we propose a full reference Stereoscopic Video Quality Assessment (SVQA) algorithm based on the Stereo Just-Noticeable Difference (SJND) model. Firstly, SJND mimic the human binocular visual system characteristics from four factors, including: sensitivity of luminance contrast, spatial masking, temporal masking and binocular masking. Secondly, based on the SJND model, the full reference SVQA is developed, by capturing spatio-temporal distortions and binocular perceptions. Finally, experimental results have demonstrated that the proposed SVQA outperforms other four current evaluation methods and has a good consistency with the observers´ subjective perception.