Genetic Programming and the Evolution of Mona Lisa

Roger Als­ing used a genetic algo­rithm to cre­ate a bril­liant approx­i­ma­tion of da Vinci’s Mona Lisa using only 50 semi-transparent poly­gons, evolv­ing over approx­i­mately a mil­lion gen­er­a­tions.

You can see the end result, after 904,314 gen­er­a­tions here, but even after roughly 100,000 gen­er­a­tions the image is impres­sive. I loved scrolling through the pic­tures, slowly see­ing the fin­ished arti­cle appearing.