A novel methodology for mapping objective video quality metrics to the subjective MOS scale

摘要

With the rapid growth in video-based services, and as users are becoming increasingly quality-aware, the reliable estimation of video quality has become extremely important. While a multitude of objective Video Quality Assessment (VQA) metrics with various performance and complexity have been proposed, the nonlinearity of video quality and the lack of clear interpretations of the metrics make difficult to understand how the objective metric values reflect the video quality as perceived subjectively in terms of Mean Opinion Scores (MOS). This paper proposes and evaluates a methodology for mapping objective VQA metric values to subjective MOS scores based on publicly available VQA databases. Three different databases were used for comparing the performance of various objective metrics and evaluating the proposed methodology.