Genetic algorithm making artistic approximations of images
The images shown in this video are approximated by a genetic algorithm. The algorithm starts by generating a population of images containing a recangle of random size, color and position. The images generated are then compared to the target image. The image with the best match is selected as the "parent" for the next generation of images. So with each generation, a "brush stroke" is added. At regular intervals, the maximum size of the generatable rectangles decreases and the size of the population increases to increase the accuracy of the final image.
00:00 Central Park in New York in autumn (24k Generations)
00:36 AI generated photo (20k Generations)
01:13 AI generated photo (16k Generations)
01:49 Watermarked photo (12k Generations)
02:25 Painting from Kavous Vesal (20k Generations)
03:01 Malenia and Radahn from Elden Ring; © From Software
Image sources:
1. SeanPavonePhoto/Adobe Stock
2. https://www.dreamstime.com/colored-st...
3. https://prompthero.com/prompt/f8c755d...
4. https://lenscraft.co.uk/photo-editing...
5 .https://www.pinterest.de/pin/28851872...
6. https://gamerant.com/elden-ring-lore-...