Painting Mona Lisa with Artificial Intelligence
summary
Recreates Leonardo da Vinci’s Mona Lisa by encoding images as genomes of filled triangles and evolving them with a genetic algorithm using a population (example size 50), pixel-wise RGB distance as the fitness function, crossover by mixing triangles, and mutation to alter colors or coordinates. Mutation is essential or the population converges. Result converges to a recognizable portrait after about 20,000 generations using only 150 triangles, demonstrating the method actually produces a close visual match.
AI-written summary · about
related talks
24mGenetic Algorithms in Go
19mRendering Mona Lisa by Solving the Traveling Salesman Problem
25mWhat if Picasso was a robot? Teaching art to the machine
57mWhat if Picasso was a robot? Teaching art to the machine
01h 27mImage Generation with AI
25mDecoding Algorithmic Artistry: Visual Generative AI
36mArt in the Age of Algorithms
38mMaking Art with Artificial Intelligence: Artists in Conversation