Alpha Portrait Machine Close-up Alpha Portrait Machine at Exhibition Portrait of Ilona Gaynor following our final crit

Alpha Portraits

Alpha Portraits plays out the self-foregrounding dynamics of the social web in a physical installation, exploring short-lived portraiture for the era of social media.

For much of history, portraiture has recorded those we value most in society. For centuries these idealised representations were primarily of royalty and religious leaders, before the 20th century brought pop-icons and political revolutionaries into the spotlight. Today, portraiture is dominated by the self-aggrandising ‘profile pics’ and ‘selfies’ of social media, which seem to promise everyone a similar place of public prominence.

An over-head depth camera tracks visitors as they walk through the exhibition space, looking for groups and attempting to identify the individual who might be the focus of that group by examining how its members are spatially positioned. Once identified, a second forward-facing camera captures the face of the focal ‘alpha’ individual and begins to draw a large portrait on a sheet of paper hanging prominently on the wall, behind an ornate frame. Upon completion, the portrait only remains for a brief period before the system identifies another group and scrolls the previous portrait to the floor to begin the next drawing.

Technical Details

The Alpha Portrait's software is a custom OpenFrameworks app (forked from BaristaBot) that takes group positioning from an overhead cam (via TSPS), and uses it to find a face at the focal point of the group (using OpenCV) . This alpha figure's face is then used to create a vector drawing which can be sent one-instruction–at-a-time to the Arduino firmware to be drawn.


Hardware teardown


Screengran of custom OpenFrameworks app

Thanks to: Kyle McDonald and Jamie Zigelbaum whose BaristaBot app I forked to begin work on the image processing. The Arduino firmware and serial communication model comes from Alexander Weber's Kritzler. The project was also inspired by wall drawing machines such as those by Jürg Lehni and Sandy Noble, amongst others.

For more…

View more photos of the project on Flickr…
Browse the project's code on Github…