An quality metric for 3D rendered image based on stereo saliency and structural similarity

In this paper, we propose an objective quality metric based on stereo saliency and structural similarity for the rendered image which is produced by depth image based rendering (DIBR) method. The rendered image and color image are not from the same view, which will lead to disparity between these two images, so the traditional quality metric cannot work. To deal with this problem, we use Scale-invariant feature transform (SIFT) to extract the matching characteristic points. The Structural Similarity (SSIM) is further used to assess compression and geometric distortions around these characteristic points. Additionally, we take 3D saliency map as a description of visual attention effect into account, which is used to determine the weights of the SSIM indices. Experimental results show that the proposed metric achieves a high correlation with subjective perception.