The supplementary material presented here provides some full-res images in addition to the ones used in the paper.
The video is a compressed version (DivX-5) of the one
available on the conference DVD.
eng1.png
eng2.png
eng3.png
eng4.png
Four different impressions of the engine rendered using the technique described in the paper. When moving the slider from one light source to an other, no repeated raycasting takes place. This is only done once in the beginning. The re-lighting is an image based operation and can thus be done at interactive framerates.
The materials used in the dataset illustrate the use of
frog1.png
frog1off.png
frog2.png
frog3.png
frog4.png
The frog dataset is another example using the same concepts as described above.
brain1.png
brain2.png
brain3.png
brain12.png
brain13.png
brain23o1.png
brain23.png
brain123.png
Three different intensity ranges of this dataset ('skin', brain, and bone) have been marked using three different material. The lights have been designed to have only one material appear while the other stay black. A linear combination of lights results in a mixture of visible segments. The filenames indicate single lights or mixtures.