Joint structure–texture sparse coding for quality prediction of stereoscopic images

A quality prediction method for stereoscopic images is proposed based on joint structure–texture sparse coding. The goal is to predict the perceptual quality of a stereoscopic image by solving the joint structure–texture sparse coding problem. First, structure and texture dictionaries from a training database are learnt. Then, the quality score for a testing stereoscopic image is predicted by computing left and right sparse feature similarity indexes, respectively, and combining them together. Experimental results on two 3D image-quality assessment databases demonstrate that the proposed method can achieve high consistent alignment with subjective assessment.