Added a simplistic scenario for using self-organizing maps in artistic work. The program uses an input image, each pixel's color is used as a training vector, and the result is a map of any size which tries to fully represent the color scheme used in the image. This is the most basic use of SOM's, however it is well suited when you want to summarize the color scheme used in an image. It should also be simple to add support for creating an Adobe color table (swatch files) from images.

More importantly, its a good example of one (out of many) uses for self-organizing maps.

input image, from the NOAA photo library	output color map, 32x32 entries

Some videos of a self-organizing map in action (using random colors as input vectors):
2.1MB Video With BMU's shown in white
0.7MB Video Without BMU's

The program has only been tested under MIPS/IRIX, but should be readily portable to Linux/BSD (with SDL and OpenGL), and a Win32 release (binary) will hopefully come soon.