Design and Applications of Perception-Based Mesh, Image, and Display-Related Quality Metrics
No Thumbnail Available
Date
2023
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Die Saarländische Universitäts- und Landesbibliothek (SULB)
Abstract
Computer graphics have become an integral part of our daily lives, enabling immersive experiences in movies, video games, virtual reality, and augmented reality. However, the various stages of the computer graphics pipeline, from content generation to rendering and display, present their own challenges that can reduce visual quality and thus degrade the overall experience. Perceptual metrics are crucial for evaluating visual quality. However, many existing methods have limitations in reproducing human perception accurately, as they must account for the complexities of the human visual system. This dissertation aims to tackle these issues by proposing innovative advancements across different pipeline stages. Firstly, it introduces a novel neural-based visibility metric to improve the assessment of near-threshold image distortions. Secondly, it addresses shortcomings of the mesh quality metrics, vital for enhancing the integrity of three-dimensional models. Thirdly, the dissertation focuses on optimizing the visual quality of animated content while considering display characteristics and a limited rendering budget. Finally, the work delves into the challenges specific to stereo vision in a virtual reality setting. The ultimate objective is to enable the creation of more efficient and automated designs for virtual experiences, benefiting fields like entertainment and education. Through these contributions, this research seeks to elevate the standard of visual quality in computer graphics, enriching the way we interact with virtual worlds.
Description