Jene is a lightweight, extensible package for creating artistic images using evolutionary algorithms in Java. Jene's algorithms and data structures are based on Karl Sims' classic paper, Artificial Evolution for Computer Graphics.
Two sample applications are included, one that produces images, and another that produces 3D parametric surfaces. Here are some sample results from each application:
- Includes two applications to help get started quickly
- Flexible, object-oriented design enables easy extension
- Includes an assortment of pre-made expression nodes and mutations
- Supports an adaptive model of evolution that learns from its mistakes
The basic algorithms and data structures used in this package are adapted from Karl Sims' classic paper, Artificial Evolution for Computer Graphics. Each Organism, corresponding to a single image or sculpture, is represented as a tree of nested expressions (Nodes). Selection of a subset of a population of Organisms is performed by an Evaluator to produce a new generation. A simple Evaluator, for example, might visualize each Organism by evaluating it over some input space and let the user choose the most attractive Organisms.
Once a new set of Organisms have been chosen to survive, mutation and sexual reproduction occurs to replace the lost members of the previous generation. Evaluation is then performed again on this new population, and the process is iterated until the user is satisfied with the results. Several Mutations have been included which manipulate the Node trees, but more can be easily defined and added.
To get the ant build to work correctly, you must download the VisAD java visualization library from ftp://ftp.ssec.wisc.edu/pub/visad-2.0/visad.jar and place it in the lib directory.
James Stout - Core library
Andy Loomis - Rendering and UI for 2D and 3D sample apps
Sam Ocko - Probability Whiz