DeepFace, Facebook’s facial recognition software, is now 97.35% accurate and “closely approaching human-level performance,” according to a new paper published by the social network’s AI research team. On the surface that level of reliability feels impressive, but for all its accuracy, what exactly does Facebook see when it looks at you?
Using Facebook’s proprietary data (in 2013 there was a short time when users could download their facial recognition template — users can no longer download this data), Brian House, a media artist and teacher at the Rhode Island School of Design, and student Jason Rabie took their own raw data as well as their friends’ and isolated the unique facial recognition templates for a series they call “Eternal Portraits.” Here’s how House describes the project:
Facebook uses face recognition software to identify its users in photos. This works via a ‘template’ of your facial features that is created from your profile images. These features — the distance between your eyes, the symmetry of your mouth — generally do not change over time. Unlike a photograph, which captures some ephemeral expression of who you are at a particular moment, a face recognition template forever remains your portrait. It is all possible photos, taken and untaken, by which you, or someone else, might document your life.
Without Facebook’s algorithm, the raw data is pretty much unusable, but it’s a reminder that Facebook has something even more personal than your data and friend network: a timeless and certifiably unique snapshot of who you are and always will be.
Here are a few images from House and Rabie’s series.