Anti-Faces created from Face Recognition-Algorithms

face0

face2

Schicke Spielerei mit Gesichtserkennungs-Algorithmen, aus denen Adelheid Mers und Robert Woodley Anti-Gesichter berechnen.

face1

The Anti Face program is face recognition turned upside-down. After looking at your image, it creates a face as different from yours as possible. It might change your gender, your age, your expression. The one dimension it will not work with much is skin color or skin tone. This is because we apply a lighting equalization calculation prior to performing the recognition.

Here are the technical details: Eigenfaces analysis is a statistical face recognition technique. It uses Principal Component Analysis to calculate a set of Eigenvectors, or Eigenfaces. These Eigenfaces can be thought of as ‘face ingredients’. To calibrate the model, we calculated 60 Eigenfaces on a training set of over one thousand faces. Whenever an image is uploaded, it is subjected to a subspace projection that reconstitutes it as a linear combination of these Eigenfaces. Normally at this point, a face recognition algorithm would look for the closest match in this 60-dimensional Eigenface space. However the anti-face calculation is not a face recognition algorithm. Instead, it creates a face opposite to yours, rather than finding a face similar to yours. It does this by multiplying each eigen value by -1.

Algopop: The ‘anti-face’ program by Robert Woodley and Adelheid Mers