Visual distortions in 360° images and their impact on Quality of Experience

This thesis comprehensively explores visual distortions in 360° images and their con- sequential impacts on Quality of Experience (QoE). Leveraging insights from exist- ing literature, a meticulously curated 360° Image Quality Dataset is introduced, fa- cilitating nuanced analysis of distortion impacts on QoE. A detailed subjective eval- uation involving 161 participants unravels the perceptual intricacies influenced by various distortions and image content. Employing Mean Opinion Scores (MOS) and ANOVA analysis, the study quantitatively assesses the perceptual impact of distor- tions across various types and intensity levels. The findings highlight the importance of customized image processing strategies to mitigate distortion effects. In addition, the performance of existing image quality metrics is evaluated in the context of 360- degree images, providing information on their suitability and limitations. Synthe- sizing key findings, this thesis advances understanding of image quality assessment methodologies for this growing medium, guiding the development of algorithms and optimization strategies to enhance user experience and satisfaction with visual con- tent. Index Terms—Omnidirectional image; 360°- image; Visual Distortions; Artifacts; per- ception; Annoyance; Dataset; Feature extraction; Visual attention; Regions of inter- est; Saliency; scene interpretation; Attention; Visual Perception