This Artistic Intervention 'Hacks' The Human Face
Robert Woodley and Adelheid Mers' "Face Field Project" turns maligned facial recognition technologies into sci-fi-worthy artwork.
Most days, facial recognition technology seems like just another step towards a global surveillance dystopia. Some days, however, artists intervene: The Creators Project has seen facial recognition applied to everything from cartoonish video chatting, to surveillance-evading masks, to even Michael Jackson dance moves.
Now, artists Robert Woodley and Adelheid Mers are experimenting with the software at the next level with a series called The Face Field Project, which gathers selfies from an app and uses them to learn all about what goes into a face.
The first item in the series, My Anti-Face, takes the facial recognition data gathered by the app and reverses it, creating the titular 'anti-face'— using software that quantifies the details of the face, it inverts each data point into its complete opposite. In this way, old becomes young, gender features become reversed, and even signs of sadness are transformed into happiness on the faces of the doppelgängers My Anti-Face produces.
The second project, The Face Spectrum, lines up each face in its database with its closest matches in 'face space.' The result is a glorious rainbow of similar-selfies. Since, in monochrome, race is factored out, the gradient is focused on less immediate characteristics, including age and mood, than on skin color. As of right now, the transition from one face to another tends to be jumpy, it will only get smoother as more people add their faces to the the app.
The last segment of The Face Field Project is The Face Synth, a system for designing a whole new faces from scratch, based on information gathered from the face database. Controlled by 60 sliders, each of which influences one part of the face's characteristics, Woodly refers to the synthesizer as 60-Dimensional Face Field, a space in which even small adjustments to a few sliders can result in dramatically different facial features. In this way, users can create images of people who have never existed in the past— and may never exist in the future.