A perceptual quality scale for image-based relighting of glossy surfaces


Olivier Dumont

Katholieke Universiteit Leuven

            

Vincent Masselus

Katholieke Universiteit Leuven

            

Peter Zaenen

Katholieke Universiteit Leuven

            

Johan Wagemans

Katholieke Universiteit Leuven

            

Philip Dutré

Katholieke Universiteit Leuven



Contact: Computer Graphics Research Group

Technical Report CW417, 8 p





Abstract

We present a method to obtain a perceptual quality scale, relating the number of basis images used in the image-based relighting of scenes containing glossy surfaces to their perceived quality. Our method is based on advanced psychological experiments, requiring observers to make judgments on different relit images. These relit images are constructed with cumulative subsets of basis images. From these judgments, a perceptual quality scale is derived.

We demonstrate this method for three different kinds of glossy materials. From the obtained perceptual quality scales, we can derive a threshold for the number of basis images required to relight the different types of glossy materials to perceptual perfection. Additionally, we determine an approximate threshold function for the number of required basis images to relight to perceptual perfection as a function of glossiness.



Keywords: Image-Based Techniques, Relighting



Downloads


Links