DIBR-synthesised video quality assessment by measuring geometric distortion and spatiotemporal inconsistency

摘要: Depth-image-based rendering (DIBR), as the most popular view synthesis method, is commonly used in application of multi-view and free-viewpoint videos. However, quality evaluation DIBR-synthesised videos remains largely unexplored, which may hinder development more advanced technology. With this motivation, Letter presents a new metric for Specifically, disoccluded regions are first detected based on an adaptive threshold to quantify geometric distortions. An energy-based sequence mapping strategy proposed portray spatiotemporal inconsistency by calculating first-order second-order similarities gradient magnitude domain Laplace-of-Gaussian domain, respectively. Finally, overall score generated pooling scores distortion inconsistency. Experimental results demonstrate that outperforms state-of-the-art metrics dedicated images