Network video quality assessment based on MDI

Abstract

In this paper, we explore the impact of network parameters on perceptual quality of network video. First we present an experiment to measure the network parameters of video in the presence of different network bandwidth. Then we propose a no-reference video quality assessment method based on Media delivery index (MDI). M5' model tree, a data mining method which is good for segmental linearization of single-output multi-output system, was introduced to model the assessment method and its parameters. For all experimental data examined, our method yields high Pearson Correlation (higher than 0.95) with measured mean opinion score (MOS).