• A-
  • A+

artistML (AInim)

artistML explores the possibilities of employing Artificial-Intelligence-Technologies (AIT) in arts and media productions. The project combines audio-reactive visuals with Image-to-Image Translation by Conditional Adversarial Networks [Isola et al 2016] trained on hand-drawn anima-tions by renowned artists. The integration of these technologies within artistic processes can be viewed as an artistic toolset rather than as a surrogate artist.

Artificial Intelligence as an Artists Tool

Live performances are often accompanied by visual elements that interact directly with artists on stage. For certain venues, an expressive and personal style is more appreciated than a digital look. However, it is impossible to achieve the expressiveness of a hand-drawn animation within a live context due to the high production effort and time requirements. Pre-produced content, on the other hand, forces the artists to perform in a given time frame and limits their artistic freedom. The project AInim delivers a solution to combine the dynamic synergy between sound and motion obtained by reactive computer graphics and the analog expressiveness of hand-drawn animations. The Ainim prototype was build in collaboration with Irina Rubina. Her animation-short "JazzOrgie" served as the foundation to set up a flexible audio reactive environment that is able to transfer the original artistic intent to a system that allows user interaction and can be entirely controlled by the artist.

Dr. Alexander König: Project-Concept, technological implementation, AI trained through hand-drawn animations by Irina Rubina.