英文摘要：
With the rapid development of mobile multimedia technology, digital media contents have become ubiquitous and are flooding and enriching our daily lives. Digital images can be easily and instantly obtained and shared through mobile phones and other smart devices. However, it is inevitable that procedures such as image acquisition, compression, transmission and restoration in digital media will contribute to distortion of original images with high resolutions, thus rendering media experience quality assessment to monitor and evaluate the effect of processing in media particularly important. At the same time, since humans are the ultimate receivers of visual signals, effective and accurate subjective experience quality of media is able to provide reliable standards for the testing, training and evaluating of objective quality assessment algorithms. At present, there are a number of image quality assessment (IQA) databases that provide subjective scores on distorted images, which use mean opinion score (MOS) and differential mean opinion score (DMOS) as evaluation indicators. Nevertheless, only using an average score to reflect the subjective overall quality of media content is lack of rationality and accuracy, which should be considered more comprehensively. For instance, the media subjective experience ratings are subject to a certain distribution, so it is necessary to improve the existing subjective quality evaluation methods. In this paper, we construct a new media experience quality assessment database based on materials of the LIVE database. Also, we investigate the relationship between statistics of subjective ratings and the type of score distribution, categories of image distortion and image content. Furthermore, this work provides a reference for further research on the subjective experience quality of media.
|