Background

In the spring of '06 I wrote a paper on using Chaos Game Representation to visualize sequences in music as a 2D image. The image makes visible the occurrence of arbitrary-length sequences. The inverse process allows one to use pixel densities to generate a markov model of the music. The order of the markov model is limited by the resolution of the image and the amount of data used to generate it.

CGRmusic.pdf is the paper I wrote.

This is a similar paper (Seems a bit more in depth, if less mathematically rigorous)

There are lots of references using CGR to visualize DNA, because the algorithm is trivially easy with a 4-letter alphabet.