This page was printed from https://textiletechsource.com

Apparel developed to confuse CCTV recognition algorithms

Out There | December 20, 2021 | By:

A German consortium has developed apparel capable of confusing AI algorithms used to track human behavior. Photo: WertelOberfell.

German design studio WertelOberfell, in collaboration with a consortium of technical partners, sought to find a way to protect people from a very modern hazard: CCTVs that are equipped with AI human-recognition algorithms. With today’s sophisticated technology, CCTV doesn’t just record video footage, it also uses artificial intelligence with human detection and facial recognition algorithms to track our behavior and even our emotions. 

The resulting prototype of their efforts is a poncho named “Ignotum” (Latin for “the unknown”) that is capable of confusing AI algorithms, essentially helping to make the person wearing it invisible to the virtual domain. 

It was designed as part of the Re-FREAM consortium, a Horizon 2020 project funded by the EU, which is intended to connect people from different fields and backgrounds to work together on combining fashion and technology. The consortium worked with technical partners from the Fraunhofer IZM, Stratasys, Profactor and Empa.

In recent years, CCTV cameras have been installed in more retail spaces to assess guests’ shopping experiences, the retail establishments say. These systems typically work in a three-step process that first filter out humans from other objects; then apply a virtual skeleton to track movement; and finally send facial images to a server for analysis.

After understanding how the AI works, the Ignotum developers rigged their own AI-enhanced camera system to find out which patterns would go undetected and found that a glowing grid pattern could confuse the CCTV computer, indicating only “33 percent person.”

Share this Story