Visual Information Vases: Towards a Framework for Transmedia Creative Inspiration

Inspiration is an important aspect of human creativity and one that creative systems are only recently implementing. In this research, we describe and implement a transmedia creative inspiration model for generative art systems. Our implementation of this model is Visual Information Vases (VIV), an artificially intelligent ceramicist that creates 3D-printable vases using inspiration from a user-supplied image. VIV scores an image along four aesthetic measures—activity, warmth, weight, and hardness—by evaluating the image’s color palette. VIV then attempts to create a vase with similar aesthetic measures through evolution. The resulting vases are diverse and functional creations. We hope that this model will allow future generative systems to create inspired artifacts from a wide variety of sources.