Objective quality assessment of stereo images with Human Visual characteristic

Quality assessment of stereoscopic images is difficult and rare because of its huge data and dependence on the development of Human Visual System (HVS). However, in this paper a novel objective quality assessment is proposed to assess the quality of stereoscopic images based on the Weighted Mean Structure Similarity (WMSSIM). A metric is predicted based on this model with cluster analysis, which provides insight into absolute disparity information of 3D images. The model can be implemented with great efficiency and speed while simulation results show that the presented objective assessment method is consistent with the subjective quality assessment and is a robust measure.