Frustrating Facial Recognition Featured Pattern: P1419 November 2019

Author: Madeeha Uppal (Send us feedback.)

Facial-recognition technologies are advancing, and techniques to subvert them are developing.

Abstracts in this Pattern:

Surveillance-camera use by law-enforcement agencies and private-security firms is on the rise and becoming increasingly advanced as object-recognition software and facial-recognition software see incorporation into camera systems. Concerns about privacy and about discrimination resulting from biased facial-recognition technology are growing. For example, during the recent protests in Hong Kong, China, demonstrators attempted to prevent the government's extensive security-camera network from identifying them by using laser pointers to thwart facial-recognition cameras.

Entrepreneurs are crafting innovative creations to deter facial-recognition technologies. For example, artist Ewa Nowak has created face jewelry that consists of two brass circles that cover the cheekbones and a long piece of brass that covers the center of the forehead. A wire that hangs over the wearer's ears like eyeglasses do connects the brass elements. Some systems use machine learning to detect humans in recordings or live feeds and then trigger facial recognition. Scientists at the Catholic University of Leuven (Leuven, Belgium) created a 2D image that can trick such systems into classifying a person as an object, thereby preventing the systems from initiating a facial-recognition check. For the 2D image to prevent human-detection software from recognizing the presence of a human, it "must be placed around the middle of a person's 'detection box' and must face the surveillance camera at all times."

Designers are also starting to create clothing that confounds surveillance cameras and protects wearers' privacy. Designer Kate Rose created a line of antisurveillance clothing that includes garments featuring a pattern of license-plate images. Automatic license-plate-reader systems see the garments as collections of license plates and add those license-plate numbers to its database, making "deploying that sort of surveillance less effective, more expensive, and harder to use without human oversight." Project Kovr (www.projectkovr.com) created a coat with a hood comprising "metalliferous fabrics" that hinder facial-recognition technologies. And researchers at Fudan University (Shanghai, China) modified a baseball cap with infrared light-emitting diodes that project spots of light onto the wearer's face to trick facial-recognition software.