Framework for the integrated video quality assessment

Mu Mu, Piotr Romaniak, Andreas Mauthe, Mikołaj Leszczuk, Lucjan Janowski, Eduardo Cerqueira

Research output: Contribution to JournalArticlepeer-review

Abstract

Through years of development Content Networks (CN) have become more sophisticated and more technically diverse. Modern CN are designed to be more adaptive to communication environment, devices and user requirements. However, one open issue is the still fluctuating quality of service provision. As a result user experience can be negatively affected. In order to maintain a satisfactory level of user experience it is crucial to develop a feasible solution to measure the extent to which video services meet users’ expectation. Assessing video quality with respect to users’ subjective opinions is a complex task. In this paper we address challenges of this task and design an integrated framework using a number of comprehensive functional modules. Our framework integrates objective quality assessment models of Artifacts Measurement (AM) and Quality of Delivery (QoD) approaches. Only the fittest models are activated by the framework considering requirements of individual evaluation tasks. We also introduce our recent work of realising key functional modules of the framework. Joint subjective experiments between two institutes have also been carried out for the purpose of model implementation and evaluation. Results from experiments verify the concept of an integrated framework and show the effectiveness of its key modules in estimating the quality level of video services.
Original languageEnglish
Pages (from-to)787-817
Number of pages31
JournalMultimedia Tools and Applications
Volume61
Issue number3
DOIs
Publication statusPublished - Dec 2012

Keywords

  • Content distribution network
  • Mean opinion score
  • Perceptual quality
  • Quality of experience
  • Video quality metrics

Fingerprint Dive into the research topics of 'Framework for the integrated video quality assessment'. Together they form a unique fingerprint.

Cite this