[Back]

T. Kment:
"Spectral data representation techniques for realistic image synthesis";
Supervisor: W. Purgathofer, A. Weidlich; Institut für Computergraphik und Algorithmen, 2011.

Digital image synthesis is used in many industrial applications. One goal of this discipline is the physical correct simulation of natural phenomena such as the color of the blue sky or refraction in translucent materials. To accomplish this goal the usage of spectral data for the representation of color information is indispensable. For this purpose different methods were developed over the years. These methods face the challenge of accurately representing complex data (e.g. fluorescent light sources). My work compares these methods using the color difference formula CIEDE2000 as well as a large data set for testing (NCS atlas), and answers the question whether the application of low pass filters during sampling rate conversions does, or does not influence the accuracy of the representation methods.

http://publik.tuwien.ac.at/files/PubDat_201071.pdf