A Neuro-Genetic Hybrid Motif Generator for Genetic Art
This work uses a physiologically inspired pulse-coupled neural network to find salient regions in an underlying image, and by subsequently introducing a motif function into the genetic programming system, is able to augment the paradigm to introduce thematic image regions.
Abstract
There have been a variety of methods for producing evolutionary art using Genetic Programming and other genetic algorithms. While some have included an underlying image, many of these systems produce aesthetically pleasing abstract images without overt structure. By using a physiologically inspired pulse-coupled neural network to find salient regions in an underlying image, and by subsequently introducing a motif function into the genetic programming system, we are able to augment the paradigm to introduce thematic image regions.