Analyses And Researches On HEVC-Oriented Video Quality Evaluation

With the diversification of electronic products, recent years havewitnessed the proliferation of multimedia content presented on varioustypes of terminals in our routine lives. As the representative of multimediacontent, digital images and video were presented to human observers tomeet the requirements of the audiovisual enjoyment of the people aftershooting, compression and transmission processing. With the developmentof video coding and communication transmission technology, consumers’requirements for video quality of services, such as IPTV, are increasing.Researching video quality assessment technologies to improve IPTV videoquality and ensure a good quality of experience is a key to gain an edge incompeting with traditional TV business. In addition, the new codingtechniques (such as HEVC coding standards) and high resolution video(such as4K video) injected new vitality for continuously improving IPTVvideo quality in the future.Starting from the actual target to improve IPTV video quality, thisarticle analyzes main factors affecting IPTV video quality based onsubjective and objective video quality evaluation studies with standard andscientific methods. Measures for each factor were made so that the videoquality of IPTV has a large progress. Then the coding efficiency of HEVCis analyzed with objective quality assessment methods. Enlightened by theresult that the coding gains obtained with different objective qualityassessment algorithm are inconsistent, the performance of objective qualityassessment algorithm are compared and analyzed from coding gainperspective. Experiments show that performance evaluation result fromSSIM is the closest to the subjective evaluation result. It provides the basis to select the appropriate video quality assessment method for accuratelyevaluating coding efficiency, and also guides on rate-distortionoptimization.Subjective video quality assessment for4K Ultra HD video encodedby HEVC has been conducted. The results provide the basis to determinebitrate in planning and layout of4K video system in the future. In addition,as a benchmark for studying objective quality assessment methods andtheir performance, the subjective result promots the development ofobjective metric for4Kvideo. Finally, considering that the large databrought by4K, full-reference quality assessment methods is difficult toimplement in practice. Moreover, researches for4K video quality mostlyfocus on high quality region. On the basis of subjective quality evaluationresults, a4K video quality estimation model based on the space-timeinformation and bit rate was proposed, which considers the impact of videocontent on the video quality. Compared to the full reference qualityevaluation methods, the model is not the best one, but it can be employedto predict video quality quickly and easily with a few video parameters.After the performance analysis we found that the model has a betterprediction performance in high quality regional than low quality region.