Sign In

Communications of the ACM

ACM TechNews

Academics Hide Humans From Surveillance Cameras With 2D Prints

View as: Print Mobile App Share:
Cameras identify one student, while the other remains "hidden."

Academics from a Belgian university have devised a method that uses a simple image that can be printed on clothing or accessories to make wearers invisible to camera surveillance systems that rely on machine learning to recognize humans in live video feed

Credit: Thys et al.

Researchers at Catholic University in Leuven, Belgium, have developed a technique of using simple two-dimensional (2D) images to render wearers invisible to surveillance cameras.

The cameras rely on machine learning to identify humans in live video feeds.

The research team experimented with various types of images, but found out that photos of random objects that go through multiple image processing operations were best at tricking machine learning systems into misclassifying a human as something else.

The same system can be modified to mask certain objects from view, if the surveillance system is configured to spot certain objects instead of people.

From ZDNet
View Full Article


Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


No entries found