Created “invisibility cloak” for cameras with facial recognition. How it works? (Video)

In recent years, the technology used in complex mass surveillance systems has achieved tremendous success. In some parts of the world, you will not even be able to go outside unless an artificial intelligence camera acts on you.

A new study, however, can easily help us confuse the video camera. A group of engineers from the University of Levinsky in Belgium created a colorful picture that you can print yourself and hang around your neck to become invisible to automatic surveillance cameras that use object recognition software based on artificial intelligence.

The preprint of the study was published last week in Arxiv.

The so-called “competitive patch” was able to hide the subjects from automatic surveillance cameras and security systems due to the fact that the image itself “effectively reduces the accuracy of human detection” – in other words, you merge with the environment, make noise and the camera cannot recognize you as a specific object.

This is not the only adversarial patch of its kind. Dutch artist and designer Simon S. Nickell created a series of T-shirts, covered with a bunch of fancy faces that can confuse Facebook software for automatic face recognition.

Dutch researchers also hope that their patch can be turned into a T-shirt, which will make users “practically invisible to automatic surveillance cameras” – at least until security system manufacturers release software updates.

Leave a Comment

Your email address will not be published. Required fields are marked *