Research On Image Quality Assessment Method For Stereoscopic Video System

Stereoscopic video system gives us stereoscopic percetion and real scenes feeling. The market demadation for development and application of stereoscopic video system becomes more and more urgent, such as stereoscopic digital televison, remote education, stereoscopic video conferencing systems, vitual reality systems and other fields, in which it all has broad application prospects. The images quality of the stereoscopic video system is related to the performance of the system, so image quality evaluation method for three-dimensional video system is the research focus currently. Through researching and analyzing human visual system and based on the research of image quality assessment, this topic researches the subjective percepton of stereoscopic images and designs objective evaluation model of stereoscopic image quality based on characteristics of human vision.First, there is the introduction to the classification of objective image quality evaluation methods, including full reference, reduced reference and no reference image quality assessment methods. In each category of evaluation methods, all are given their research ideas, advantages and disadvantages, as well as detail introducetion to typical models of each evaluation. The intoduction to evaluation method is focused on quality assessment of blockiness artifacts, based which a blockiness artifacts based on human visual system is proposed according to the human eyes'different sensitivity to the flat and edge regions and combination of luminance and texture masking. Experimental results illustrate that the proposed metric has good consistency to the visual perception, and it can reflect the subjective perception of image quality correctly.Second, Research on symmetric stereoscopic masking for stereoscopic video coding. Aiming at requiring the masking of stereoscopic images in symmetric stereoscopic video coding, paired comparison is used for subjective perception experiment. Through analyzing and modeling the results of subjective perception experiment, the regularity of symmetric stereoscopic masking in stereoscopic perception is obtained. The results show that the stereoscopic images quality threshold point is related to the quality and content of stereoscopic images. The subjective perception results provide experimental basis for the rate allocation and are instructive to select parameters in the symmetric stereoscopic video coding.Finally, based on objective evaluation methods of image quality assessment and subjective perceptual characteristics of stereoscopic images, combined with human visual properties, a quality assessment metric of stereoscopic images is proposed. Through simulating contrast sensitivity function, multi-channel effects and stereoscopic perception of human visual system, the metric makes full use of contrast sensitivity function, multi-channel effects and stereoscopic perception of human visual system, which contains evaluation of image quality and stereoscopic perception. Then regression analysis was used to integrate the two evaluation results into an equation as our quality assessment model of stereoscopic images. Experimental results show that the proposed objective metric achieves consistent stereoscopic image quality evaluation results with subjective assessment.