This page contains datasets and supplementary materials for the paper.
The github repository providing the implementation of the light source estimation is available here.
We forked the 1st version of Mitsuba to render hyperspectral-like hypercubes and RGB images of a scene.
There are two different branches which need to cloned, compiled and built:
These repositories should be in separate directories. The spectral version renders hypercubes of 100 channels.
The scene files to render examples from the paper are found in here. The README will describe the process to generate datasets.
The supplementary material for the paper is available here.