Computational Models of Abstract Art (2018)

During fall of 2018, Williams College students in my Mathematical Modeling course created computational models of abstract art. This project would not have been possible without the support of the Williams College Museum of Art, and especially, Liz Gallerani. Use the links below or scroll down to see the assignment prompt, and under that, original works and works inspired by them, created algorithmically.

Assignment prompt
Barbara Takenaga, Roof Fuji
Abraham Walkowitz, Abstract Composition
Anni Albers, TR II
Stuart Davis, Study for Allee
James Rosenquist, Circles of Confusion

The first time I saw Barbara Takenaga’s work, I was immediately taken because I, as a mathematician, found it to be so mathematical. In this modeling exercise, you will study abstract works of art by Takenaga and others.

In your group, please model your assigned work by creating an algorithm that creates one or more images similar to it. Your goal is not to exactly reproduce the work, but rather, to create a work that is in the same style. Once you create your art, make sure to compare and contrast it with the art that inspired it.

There are many ways to interpret any piece of art! When I view abstract art through a modeling lens, I am tempted to think about distributions and degrees of randomness. What sorts of shapes are there? Are there variations? How are these shapes distributed: regularly or irregularly? Is color constant or are a range of colors used? How do shape and form vary across the canvas?

When creating your model, draw from the work under consideration as quantitatively as possible. This means that you may wish to do some analysis of the original piece of art, which is why I have provided you the digital files. There are many computational tools to help you, including (but certainly not limited to) the Matlab Image Processing Toolbox.

In mathematical modeling, we sometimes say “all models are lies, but some models are useful.” That is to say, any model is necessarily a simplification of real life, and yet, sometimes models capture certain key features of reality. While no algorithmic model will replace artists like those we are studying, I hope that modeling a work of art will prompt you to think about artistic process at a very granular level.

Present your results in a recorded presentation, submitted online, following these guidelines. Your presentation is due on the date indicated in the course schedule.

Barbara Takenaga, Roof Fuji