Rendering Realistic AR Content

A VEIV EngD Project by David Walton, 2013 cohort

The focus of this EngD project is on the rendering of realistic virtual content in augmented reality and mixed reality applications. Augmented reality and mixed reality applications attempt to add virtual content to the real world, for example by augmenting the output of a camera in a mobile phone application, or by displaying content on a translucent display such as the one used in the Microsoft Hololens.

In many cases, the goal of such an application is to add virtual content which is indistinguishable from the real-world content. This presents a wide array of challenges. In addition to the computer graphics challenges of rendering realistic content at interactive framerates, there is the added difficulty associated with obtaining the necessary information from the real environment, and estimating the influence of the virtual content on the real world.

This project aims at developing novel ways to collect this information from the real environment, and use it to render more compelling virtual content in real time. It aims to build on existing approaches by removing the need for external light probes, and adding the ability to respond to rapid changes in real-world geometry and lighting.