Saliency structure stereoscopic image quality assessment method

Stereoscopic image quality assessment (SIQA) plays a crucial role in the development of 3D imaging system. In this paper, an objective SIQA model named saliency structure stereoscopic image quality assessment method (3SIQA) is built, based on the fact that human is selectively sensitive to various structure. Structural similarity image index (SSIM) is generally used to calculate the structure similarity between the reference image and the distorted image, and the selective sensitivity of human vision system is described on two aspects: saliency and texture. The former aspect performs through giving different weights to the image pixel value and its saliency map, and the latter one is represented by dividing SSIM map into three parts (edge, smooth and texture zone) and summing these parts with different weights. Moreover, the experimental results demonstrate its promise.