<--back--      home      --next-->

This paper demonstrates what I think is an interesting point. As far as these colormaps are concerned, the property of a colorspace being "perceptual", and being "device-independent", are orthogonal. We did perceptual colormap generation in a completely device dependent space.

The obvious next step for this work is to do some other methods to make the colormaps perceptually uniform. Some of the methods described in the very nice Psychometrics 101 half-day course on Tuesday could perhaps be used for this. Lefkowitz has described some methods for doing this which employ the CIE perceptually uniform, and device independent, colorspaces.

Also I want to mention some problems with the final version of the paper. Basically, the explanations aren't very good-- my understanding of the problem has improved, and other people have helped me find better ways of explaining what we did. Also, making the rainbow colormap isoluminant may not have been the best task for the user study. By varying colors to match a fixed gray, instead of vice versa, the saturation was reduced on most of the colors, which means that it was a less effective comparison against MDB, as far as sensitivity to how the Helmholtz-Kohlrousch effect could confound the matching method.

The sofware which I demonstrated during the demo should be available very soon, its just a little VTK program.