Constant Subjective Quality Database: The Research and Device of Generating Video Sequences of Constant Quality

Abstract

Video databases with the collection of subjective assessments form an important basis for learning algorithms, testing, and creating an objective video quality assessment. In this paper, presented is a novel device to measure coded video quality was proposed using the finding of an acceptable minimum perception threshold, which allows one to generate video sequences of constant quality. Also, this work describes the method, which allows generating video sequences of constant quality, and a specialized installation that allows you to dynamically change the visual quality of typical media content during the examination process. This approach will allow obtaining well-labelled data in a larger volume compared to the currently used binary estimates. The novel databases with this approach will be used both to create effective adaptive codecs built according to the hybrid model, and new ones based on machine learning.