Project R-14292

Expanding Capabilities of View-Dependent Lighting Effects in XR (Research)

EXtended Reality(XR) has become a popular topic over the last years as major tech companies are investing heavily in improvements to both XR software and hardware. The realistic content for these devices is currently limited as both creation and rendering can prove to be difficult. The past two years, significant progress has been made concerning the topic of lighting effects as the seminal Neural Radiance Fields (NeRF) and Plenoxels papers have shown great results in this domain. Both papers however, have difficulties presenting high frequency details which can lead to unwanted results. Especially for users, it is important to show highly detailed content as small changes from reality can result in nausea and a general unpleasant experience. My proposal will focus on achieving this level of detail in an efficient manner so it can be used for XR applications. During this research the focus lies firstly on improving the amount of detail for view-dependent lighting effects and secondly on improving detail overall. Using both of these approaches I can continue my research by making it possible to interact with objects in a scene while maintaining realistic lighting effects on moving objects. As efficiency is critical in these applications, I will research sufficiently efficient approaches which comply with the restrictions XR puts on these applications such as a high enough framerate.

01 November 2023 - 31 October 2025