The average Londoner is seen by CCTV cameras more than 300 times a day - that is, in the British capital, there is one camera for every 14 people. This figure is projected to rise to 1-for-11 over the next five years. Taking into account the trend, researchers at the Massachusetts Institute of Technology and IBM have developed a T-shirt that makes a person invisible to facial recognition systems.
Conventional recognition algorithms work like this: the system "sees" a characteristic image, draws a "bounding box" around it, and assigns an appropriate identifier to this object. To break the chain, the engineers placed vibrant pixel patterns on the shirt. The drawing is structured to confuse the artificial intelligence-based classification and labeling system. This makes it difficult for her to identify the facial features of a person.
Experiments have shown that wearing such a T-shirt reduces the likelihood of being detected by digital surveillance technology by 63%. Although such cyber clothing still has difficulties with implementation outside of laboratories and requires improvements.
Today, to confuse facial recognition algorithms, people are resorting to cosmetics. It takes a lot of paint and time for a make-up to hide key facial features such as the eyes or the bridge of the nose. In the long term, such laborious methods may be replaced by just simple and unpretentious clothing with the effect of “digital camouflage”.