The objective quality assessment of stereo image

The target of stereo image quality assessment is to establish a computational model so as to predict the visual quality of a stereo image. A few new depth map methods have been proposed for the last few years. However, these methods are not suitable for all stereo images. Since the depth map fails to capture the directional reflection of human stereo sense, it is not easy to compute depth map accurately. In this paper, a new quality assessment method of evaluating the performance of stereo images is proposed. Differing from the current depth map methods, this method focuses on a new aspect. It assesses stereo images from the perspective of image quality and stereo sense based on objective quality assessment of 2D image and absolute disparity image. The theoretical argument is supported by experimental results implemented on a series of benchmark data sets.