Optimize Evaluation Of Stereo Video Comfort Based On Luminance

Optimizing the evaluation of stereo video visual comfort based on luminancebelongs to the stereo video visual comfort evaluation technology. With the rapiddevelopment of the stereo display technology, the video display technology producesa significant change. From black and white to colour, from analog to digital, fromstandard definition to high definition, to stereoscopic video now. In the continuousresearch of former people, stereoscopic video technology is gradually mature. InOctober2011, a new type of stereo video display technology is listed as one ofShanghai "twelfth five-year" plan key development fields of science and technology,Shanghai science and technology development research center noted that3Dtelevision market will become the biggest application market in our country in thefuture, so it is very urgent to the research of stereo video.As the stereo video now stereo video display technology is still not perfect, andthe content is not compatible, and the viewing environment is very limited, there oftenappears some visual physiology abnormal phenomenon in the process of stereo videoviewing experience, such as dizziness, vertigo, nausea, or vomiting. Visual comfort ofwatching stereoscopic video has direct relation to the viewer’s physical health, andthis is a major factor restricting the development of3D products. This view isgradually identified by a wide range of stereo video users. So far, the world of stereovideo visual comfort evaluation does not have unified standard, therefore, to study theinfluence factors of stereo video visual comfort is very important, and how to set up astereo video based on human visual system visual comfort evaluation model also hasan important significance.In order to improve the previous visual discomfort model, this paper establishesa stereoscopic video visual comfort evaluation model which is more consistent withhuman subjective perception results, and whose score is closer to the human visualsystem characteristics of subjective test score. The method is: extracting theforeground region of stereoscopic video combining the grayscales of left and rightview, and then synthesize the two pictures into a disparity map, extract the horizontalparallax, and calculate the depth parallax. Finally， build the depth parallax-visualdiscomfort model on the basis of subject assessment score and optimize the visual discomfort model by considering the luminance influence. Experimental resultsindicate that the error rate is less than5%between calculated value and humansubjective evaluation results, and most video error rate are less than1%. So, thecomfort model is the optimization of predecessors’ visual comfort evaluation model,and the evaluation model is more accord with human subjective perception results.The model has good practicability, and promotes the development of stereo videocomfort evaluation.