This post is also available in: heעברית (Hebrew)

A new project aims to reduce the confidence score of facial detection and recognition by providing false faces that distract computer vision algorithms.

A Berlin-based artist has developed a concept that borrows from the camouflage techniques used by animals.
With the HyperFace project, Adam Harvey uses thousands of “algorithm-specific optimized false faces which reduce the confidence score of your true face.”

These patterns, which can be printed on clothing or textiles, appear to have all facial features that the visual software can interpret as a face – patterns of eyes, noses, and mouths – thus overloading and over saturating the algorithm so that it can’t really tell which faces are real, according to The complex patterns may be worn directly or used to flood an area to confuse security cameras.
The technical concept is an extension of Harvey’s earlier project, CV Dazzle, in which he used avant-garde hairstyling and makeup designs to break the continuity of a face — a so-called “anti-face” — and prevent facial-recognition software from detecting it.
According to the project’s site, HyperFace works by providing maximally activated false faces based on ideal algorithmic representations of a human face. The difference between the two projects is that HyperFace aims to alter the surrounding area (ground) while CV Dazzle targets the facial area (figure). In camouflage, the objective is often to minimize the difference between figure and ground. HyperFace reduces the confidence score of the true face (figure) by redirecting more attention to the nearby false face regions (ground).

Conceptually, HyperFace recognizes that completely concealing a face to facial detection algorithms remains a technical and aesthetic challenge.

“I got inspiration from false coloration in the animal kingdom,” Harvey said. “HyperFace is about reimagining the figure-ground relationship of the human body to our environment in the context of computer vision.”