Quality Assessment of Stereoscopic Video in Free Viewpoint Video System

Abstract It is challenging to develop an effective quality assessment method for the stereoscopic video (SV) in the free viewpoint video (FVV) system because of asymmetric distortion and the interactively changing combinations. This paper proposes a quality assessment method that takes the characteristics of the SV in FVV systems into consideration. Specifically, considering the distortion introduced by the rendering process, the proposed method extracts a critical distortion area that can be perceivable by the human eye. The distortion degree of each critical distortion area is quantified and the overall results are pooled to obtain the initial quality score. The video pairs are classified into different combination types and the initial quality score is refined to obtain the final quality score. The experimental results show that the proposed method outperforms conventional objective methods.