Objective quality assessment for stereoscopic images based on structure-texture decomposition

We present a novel quality assessment index for stereoscopic images based on structure-texture decomposition. To be more specific, we decompose a stereoscopic image pair into its structure and texture components. Then, gradient magnitude similarity (GMS) and luminance-contrast similarity (LCS) indexes are used to measure the qualities of structure components and texture components, respectively. Finally, the quality score is obtained by combining the above quality scores in a non-fixed manner. Experimental results on two publicly available 3D image quality assessment databases demonstrate that, in comparison with the related existing methods, the proposed technique achieves high consistency alignment with subjective assessment. Key-Words: Quality assessment; structure-texture decomposition; gradient magnitude similarity; luminancecontrast similarity.