Mathematicians Reimagined

Abstract

I was fooling around with AI image generators and came across stable diffusion models and extensions that allowed me to generate modern images of famous mathematicians using old existing public domain images as input. I present here a few of my results. The people displayed here are sorted by birth year.

Euclid (Sometime in ~350-200BC) reimagined.

René Descartes (1596-1650) reimagined.

Issac Newton (1642-1727) reimagined.

Gottfried Wilhelm Leibniz (1646-1716) reimagined.

Leonhard Euler (1707-1783) reimagined.

Pierre-Simon Laplace (1749-1827) reimagined.

Carl Friedrich Gauss (1777-1855) reimagined in a modern look.

Carl Gustav Jacob Jacobi (1804-1851) reimagined. Fun fact: Jacobi is not living in my neighborhood.

Bernhard Riemann (1826-1866) reimagined.

Georg Cantor (1845-1918) reimagined.

Henri Poincaré (1854-1912) reimagined. He became a bit of a hipster.

David Hilbert (1862-1943) reimagined.

Emmy Noether (1882-1935) reimagined.

Workflow

• Stable diffusion 1.5 img2img
• ControlNet with various settings
• ReActor face swapping
• Lots and lots of fixes

I can't really be more specific without going into endless explanations when to use which ControlNet model with what settings. No two examples here were made with the same pipeline. I had to test my way around to get images I am satisfied with. Many ideas were not realized, especially regarding hair and beard features.

👉 Email me if you have a specific request for a mathematician.

Bonus Content

Using a few more AI tools, I managed to recreate more creative stuff.

More Euler images. Left: maintaining the painting style (weird hat). Right: Leon-Hard Euler.

About Copyright

I personally make no copyright claims on this output and advise you to be careful regarding copyrights of these images, as they are output from AI models trained on unknown datasets. You can still cite this webpage using the BibTeX below.

BibTeX