Now showing 1 - 1 of 1
Thumbnail Image

Through the Looking Glass: Working Towards the Ultimate 3D Display by Adapting the Light Field Displays using Context-Awareness

2022, Pan, Xingyu, 0000-0002-5585-1975

Ivan Sutherland described an ``Ultimate Display" as a true ``window" that connects the computer-generated virtual world with reality. Through this window, the viewers should freely view and interact with the virtual objects. This concept was widely accepted by academia and industry. Therefore, during the past decades, creating the ultimate display has become the goal of the next generation of three-dimension (3D) displays. By reviewing the state-of-the-art 3D displays, we found the Light Field Display (LFD) is the optimal modality to achieve this ultimate display. Nevertheless, the LFDs still suffer from some drawbacks, such as the high price, lack of interaction with the users and the external world, and insufficient motion parallax. Previous research has been working on solving these problems by upgrading the hardware while rarely focusing on the software level. However, with the recent software-based methodologies such as context awareness, it becomes possible to improve the current LFDs without altering the hardware. Moreover, the software-based solution will allow the current users to obtain new features by updating the LFDs Over-The-Air (OTA), making it an economical and environmentally-friendly approach for LFD upgrading. Therefore, this thesis will focus on solving these issues and improving the off-the-shelf LFDs through context awareness approaches. To reduce the cost of the current LFDs, this thesis will illustrate an approach that lowers the Graphics Processing Unit (GPU) consumption to reduce the hardware cost. A performance evaluation proves that this approach can save GPU consumption significantly. The advantage will be further enlarged when applied to the higher resolution LFDs. Moreover, this thesis will describe a framework to multiplex one LFD for multiple viewers. With this framework, one LFD can work as multiple displays to show the viewers different 3D contents simultaneously, which further reduces the usage cost of LFDs. To bridge the interaction gaps among the LFDs, the users, and the external world, this thesis will present three collaborative systems between LFDs and other Mixed Reality (MR) headsets. These systems enhance the interaction between the LFD and viewers and enable the connections between the LFDs and reality step-by-step. With the exploration of these collaborative systems, the LFD evolves from a display modality to an interactive visualization tool. Finally, this thesis will describe an approach to provide the LFDs with additional axes of motion parallax along the y and z-axis. Furthermore, a user study will be presented to explore the effects of these additional axes of motion parallax on viewers' 3D perception. This thesis presents several context-aware adaptions and empirical results to accelerate the development of the LFDs. These technical contributions aim to make future LFDs more affordable, interactive, and immersive. The empirical findings and the design principles may guide future designers and researchers to continue improving the LFDs to the real ultimate displays. Such LFD-based ultimate displays will lead the evolution of 3D displays to widely affect people's daily lives. This research also pointed out two promising research directions, such as utilizing the context-aware LFDs as personalized advertisement billboards and visual correction displays, which are interesting and worthy to be resolved by the community in the future.