start

We collected 200 selfies on Instagram. These are selfies of people around the world who have used the hashtag #mouthmask.

We have searched for portrait photos to show the pattern as clearly as possible.

BY ILJA YAEL SPRENGER & MELISSA YANELL V/D REE

The Generative Faces project is an augmented reality exhibition that grew out of our interest in patterns that influenced our lives during Covid-19.

One of these patterns is the mouth mask, which is something everyone in society has to deal with and which should reduce the spread of Covid-19. But it also affects several things. It ensures that people hide their identity and can no longer show or read emotions. In addition to the 1,5 meters distance we have to keep from each other, the mouth mask also creates distance between people because making contact has become less easy and personal.

Normally there are cameras all around us that can recognize our face. But lately, half of our face has been covered by a mouth mask and everyone in public spaces is basically anonymous.

We uploaded our data collection of mouth mask selfies to an AI machine. The machine normally recognizes the human face and is able to mix multiple faces to get a new outcome. We wanted to see how this machine translates this pattern of algorithms, while a part of the human face is not recognized. What are the results of this, what data does the machine absorb from the human face and can we still be identified by our visible external features?

With an input of 200 photos, the machine returned an output of 3200 photos. From this, we have made a selection of the most interesting results.

The results show that the unrecognizability that was already there is only further reinforced. The machine sometimes confuses the mask with the skin. The eyes are the only thing that makes you recognize them as people. You know it's about people, but it looks very unnatural and new creatures arise, as it were.

The selection of shapes and the 3D sculpture together form the exhibition. By taking the unrecognizability and loss of identity through AI out of context, we created an exhibition that represents the reflection of society during Covid-19. How the machine handled this data from this pattern shows that there is a relationship between a technological error and the human being. Augmented reality is an ideal way to hold an exhibition during a pandemic. It also ensures that people have a reason to go outside.

We have made a 3D sculpture that represents the reflection of a generative society.

To experience this AR exhibition, you need to download Adobe Aero and scan the QR code. As a result, this exhibition can be viewed wherever you want.