Plenoptic Light Transport

Abstract

In this thesis we focus on the multidimensional nature of light transport as described by the plenoptic function, and in particular in the angular and temporal domains. While traditional imaging has been limited to bidimensional images, the emerging field of Computational Imaging has made increasingly available more complex multidimensional visual data, disambiguating additional domains of the plenoptic function. However, this higher dimensionality requires changing the way that visual information is processed, manipulated, visualized or synthesized. In this thesis we present contributions on these topics, addressing the challenges of adapting and rethinking them to handle higher-dimensional visual information. Specifically, within the angular domain we focus on light field editing, studying interaction paradigms and user workflows when interacting with light fields, and on spatio-angular filtering of complex appearances modeled with BTFs, studying how filtering affects appearance perception. On the other hand, in the temporal domain we focus on transient light transport, where the speed of light cannot longer be considered infinite, including contributions on capture and data processing, light transport simulation and visualization of time-resolved data.