Imagine the fusion of graph theory and global illumination to significantly reduce the time required to render global illumination images with physically-accurate features such as refractions, caustics, and diffuse bleeding. My master's thesis work addressed this issue, rethinking the foundational algorithm for ray-based global illumination algorithms. Useful in almost any global illumination algorithm, my algorithm converts rendering into a two-pass algorithm, which allows optimal reuse of ray paths with very little computation cost.

The images on the right give examples of what our algorithm can produce in about half the time required by traditional algorithms. My thesis work was accepted in the Computer Graphics International 2011 conference, where I presented in Ottawa, Canada in June of 2011.

Graph-Based Global Illumination