What is Open Source Art?

While traditionally used for programming, this reposity combines open source development with community artistic contributions to create the first automated, programatically and manually generated gallery.


Background

There is typically a strong distinction between things that are made by hand, and by machine. It's also more likely the case that you see artist work attributed to one or a few individuals - the paintings in your favorite gallery are done by one individual, or it's perhaps a collaboration between a small group.

Bring in open source. Open source is a powerful framework with under which many eyes can make tiny contributions to develop highly complex systems. In the context of programming, this means that a group of developers opens up their code base for contributions from the community, and many eyes on the code squash the bugs.


How does this work?

Under this same model, many small artistic contributions can create unexpected beauty.


1. The Human Component

The base of the entire gallery, whether it be code that drives its generation or the human hand, starts with a human component. How does this work?

Any member of the community can contribute a texture, a graphic, or digital piece. The contribution is done via a pull request (or contact of one of the Open Source Art maintainers, in the case the artist is not familiar with Github). Using Github means that the contribution (if necessary) can be discussed and reviewed, and then finally the contribution is added to the library housed in the repository.


2. The Digital Gallery

From the human component, we have a version controlled collection of community contributions. Each carries with it metadata about the artist and (if one exists) a title and description. Here is where the machine, and concepts from software engineering like version control, continuous integration, and machine learning come into play! We use these technologies to automate the process of generating new works from the existing.

For each submission, we use a deep learning algorithm to generate a stylized photo during an automated testing step in a continuous integration service. In addition to a few derivations, a previous submission is also selected from the gallery to generate a combined piece, meaning that the gallery grows on itself over time. When the submission is approved, all images are added programatically back to the gallery.


3. The Open Source Art Story

Over time, the gallery grows in size and complexity. Earlier pieces become part of the later ones, and the container serving the algorithm can be further developer to change how the images are generated. The entire effort, programming and art, combines the beauty of code and visual design to create a dually man and machine-made story.

How do I contribute?

So glad that you asked! Read about how to contribute here