Quality of Experience in Immersive Video Technologies

Over the last decades, several technological revolutions have impacted the television industry, such as the shifts from black & white to color and from standard to high-definition. Nevertheless, further considerable improvements can still be achieved to provide a better multimedia experience, for example with ultra-high-definition, high dynamic range & wide color gamut, or 3D. These so-called immersive technologies aim at providing better, more realistic, and emotionally stronger experiences. To measure quality of experience (QoE), subjective evaluation is the ultimate means since it relies on a pool of human subjects. However, reliable and meaningful results can only be obtained if experiments are properly designed and conducted following a strict methodology. In this thesis, we build a rigorous framework for subjective evaluation of new types of image and video content. We propose different procedures and analysis tools for measuring QoE in immersive technologies. As immersive technologies capture more information than conventional technologies, they have the ability to provide more details, enhanced depth perception, as well as better color, contrast, and brightness. To measure the impact of immersive technologies on the viewersâ QoE, we apply the proposed framework for designing experiments and analyzing collected subjectsâ ratings. We also analyze eye movements to study human visual attention during immersive content playback. Since immersive content carries more information than conventional content, efficient compression algorithms are needed for storage and transmission using existing infrastructures. To determine the required bandwidth for high-quality transmission of immersive content, we use the proposed framework to conduct meticulous evaluations of recent image and video codecs in the context of immersive technologies. Subjective evaluation is time consuming, expensive, and is not always feasible. Consequently, researchers have developed objective metrics to automatically predict quality. To measure the performance of objective metrics in assessing immersive content quality, we perform several in-depth benchmarks of state-of-the-art and commonly used objective metrics. For this aim, we use ground truth quality scores, which are collected under our subjective evaluation framework. To improve QoE, we propose different systems for stereoscopic and autostereoscopic 3D displays in particular. The proposed systems can help reducing the artifacts generated at the visualization stage, which impact picture quality, depth quality, and visual comfort. To demonstrate the effectiveness of these systems, we use the proposed framework to measure viewersâ preference between these systems and standard 2D & 3D modes. In summary, this thesis tackles the problems of measuring, predicting, and improving QoE in immersive technologies. To address these problems, we build a rigorous framework and we apply it through several in-depth investigations. We