Prof. Wojciech Jarosz, together with Romain Prévost (ETH Zurich/Disney Research Zurich), Alec Jacobson (ETH Zurich/Columbia University), and Olga Sorkin-Hornung (ETH Zurich) have developed a computationally-assisted spray painting system that allows novices to spray paint accurate reproductions of photographs as large murals. The research appears in April's issue of the journal Computers & Graphics.
The group wanted to create a way to allow untrained users to spray paint large graffiti murals. To accomplish this, they instrumented a physical spray paint can with an actuator to turn it on/off and QR codes to track its location. A nearby computer analyses the position of the spray can as well as the current painting on the wall/canvas, and automatically turns the spray on or off. In essence, the spray can "knows" what it wants to paint, and the user just has to wave it around, seeing the image appear in front of them.
For more technical details, including a preprint of the publication and a demonstration video, check out the project page on Prof. Jarosz's website.
Several news agencies are also reporting on the research:
Wall Street Journal | Spray Cans That Know What to Paint
Wall Street Journal Video | Scientists Develop Spray Cans That Know What to Paint
NeuroScienceNews | Scientists Invent Robotic ‘Artist’ That Spray Paints Giant Murals