Master’s Thesis Presentation • Computer Graphics • A Modular Framework for the Plausible Depiction of the Starburst Patterns

Please note: This master’s thesis presentation will take place in DC 3317.

Yuxiang Sun, Master’s candidate
David R. Cheriton School of Computer Science

Supervisor: Professor Gladimir Baranoski

When human viewers look at an intense light source, they can observe a set of dense spikes radiating from its center. The observed spike patterns result from a complex optical process known as the starburst phenomenon. These patterns have been frequently depicted in virtual applications (e.g., films and video games) to enhance the perception of brightness. From a broader scientific perspective, they also have relevant real life implications such as the impairing of driving safety. This may occur when an observed starburst pattern saturates a driver’s field of view.

Previous computer graphics works on the simulation of the starburst phenomenon have primarily relied on the assumption that the resulting patterns are formed by light being obscured and diffracted by particles in the observer’s eyeball during the focusing process. However, the key role played by background luminance on the perception of starburst patterns has been consistently overlooked.

By also taking into account this pivotal factor as well as the different physiological characteristics of the human photoreceptor cells, we propose a modular framework capable of producing plausible visual depictions of starburst patterns for daytime and nighttime scenes. To enhance the physical correctness and resolution of the simulated patterns, its formulation was guided by the Rayleigh-Sommerfeld diffraction theory and implemented using the Chirp Z Transform, respectively. We also introduce a biophysically-inspired algorithm to enable the seamless incorporation of the resulting patterns onto computer-rendered scenes. Besides examining theoretical and practical aspects associated with the use of the proposed framework in realistic image synthesis, we also derive biophysical insights that may contribute to the current knowledge about the inner workings of the starburst phenomenon.