Computational Inference and Control of Quality in Multimedia Services

This thesis focuses on the problem of optimizing the quality of network multimedia services. This problem spans multiple domains, from subjective perception of multimedia quality to computer networks management. The work done in this thesis approaches the problem at different levels, developing methods for modeling the subjective perception of quality based on objectively measurable parameters of the multimedia coding process as well as the transport over computer networks. The modeling of subjective perception is motivated by work done in psychophysics, while using Machine Learning techniques to map network conditions to the human perception of video services. Furthermore, the work develops models for efficient control of multimedia systems operating in dynamic networked environments with the goal of delivering optimized Quality of Experience. Overall this thesis delivers a set of methods for monitoring and optimizing the quality of multimedia services that adapt to the dynamic environment of computer networks in which they operate.

Awarded the best doctoral project award 2014 at Eindhoven University of Technology Develops a novel application of computational intelligence to address complex QoE optimization problems Provides a new QoE estimation method for video streaming services based on machine learning Includes supplementary material: sn.pub/extras

Contenu
Supervisor's Foreword.- Preface.- Introduction.- Objective QoE Models.- Subjective QoE Models.- QoE Management Framework.- QoE Active Control.- Conclusions.