Quality assessment of videos compressed by HEVC based on video content complexity

Abstract

In the last few years, online video services have experienced great development. When we watching videos, we all want to get better watching experiences, but the delivery of the videos are restricted by Internet bandwidth. In order to get better video quality with limited Internet environment, people have formulated a lot of compression standards to make the size of video files smaller. But during the procedure of video compressing, the video quality may decrease. So it is important for us to use appropriate encoding parameters for the videos with different content complexity. In 2013, the newest compression standard called High Efficiency Video Coding (HEVC) was promoted by the Joint Collaborative Team, and it is also called H.265. So far, a lot of encoders have been developed on the basis of HEVC standard and the most widely-used one is x265. As we know, video content complexity can be affected together by spatial perceptual information (SI) and temporal perceptual information (TI) of test sequences. We use scene criticality (CI) to combine the two factors. In the paper, we want to find the relationship between quality of video sequences with different resolutions and their content complexity. In the experiment, the test video sequences have diverse content complexity and they are all compressed by x265 encoder with different encoding settings.