On Graffiti, art, perception, human-machine understanding.
AutoGraff is a project that focuses on the computational generation of forms based on the style and process involved in drawing graffiti letters. We base our study on the direct experience of Daniel Berio as a graffiti artist and analyze the process he follows when sketching. This leads to the construction of generative models that abstract the main components of the process and style under analysis.
The resulting generative systems are capable of creating images that are consistent with the stylistic signature of Graffiti Art. The system becomes a familiar extension of the artist and his style into the computational medium, allowing the same stylization principles specific to graffiti letters to be applied to different types of visual media. We call this process Graffitization: The (computational) process that applies the stylistic principles of graffiti art to computer generated forms.
At its current state this study has led to the creation of artworks and interactive installations where the process of graffitization is executed in an autonomous manner by drawing machines. This allows a subsequent seamless interaction with the computer generated work by using traditional artistic media, or the augmentation of the drawing process via digital techniques as projection mapping. The works generated by the machine have a strong resemblance to the work that would be executed manually by the artist, and the stylistic signature of graffiti is well recognizable also by peer graffiti writers.