Genetic algorithm finds Mona Lisa in the lake of possibilities

Different target space: https://www.youtube.com/watch?v=nvmo1SCBDSM

The algorithm starts by creating a population of images with random RGB values for each pixel.
Each color is evaluated separately for each individual.
Thus, 10 images correspond to 30 individuals.
The individual with the best matching red values gives its red values to the new parent. Another individual that best matches the green values provides the green values, and the same is true for blue.
The second best individuals then form the other parent

These two parents breed the next population by randomly combining parts of one parent with corresponding parts of the other parent.

Finally, the offspring is randomly mutated with adaptive probability.

With each successive generation, the mutation rate must decrease to further improve quality.

This code was created in MATLAB

Music: Kevin MacLeod - Local Forecast

Видео Genetic algorithm finds Mona Lisa in the lake of possibilities канала Hali Hammer