CAN: Creative Adversarial Networks, Generating "Art" by Learning About Styles and Deviating from Style Norms

TL;DR

CAN introduces a novel GAN-based system that generates creative art by learning styles and intentionally deviating from them, producing images indistinguishable from human-created art and often rated higher by viewers.

Contribution

The paper presents a modified GAN framework capable of generating creative art through style deviation, advancing computational creativity in art generation.

Findings

Humans could not distinguish generated art from artist-created art.

Generated art was rated higher on aesthetic scales.

The system successfully produces creative and novel images.

Abstract

We propose a new system for generating art. The system generates art by looking at art and learning about style; and becomes creative by increasing the arousal potential of the generated art by deviating from the learned styles. We build over Generative Adversarial Networks (GAN), which have shown the ability to learn to generate novel images simulating a given distribution. We argue that such networks are limited in their ability to generate creative products in their original design. We propose modifications to its objective to make it capable of generating creative art by maximizing deviation from established styles and minimizing deviation from art distribution. We conducted experiments to compare the response of human subjects to the generated art with their response to art created by artists. The results show that human subjects could not distinguish art generated by the proposed system from art generated by contemporary artists and shown in top art fairs. Human subjects even rated the generated images higher on various scales.