Invisible sweater: scientists created a special print on clothing that "breaks" AI systems of human recognition
Staff and students at the University of Maryland have developed a special sweater that will help trick AI cameras into recognizing people.
Here's How It Works
The developers' original goal was to test machine learning systems for vulnerabilities. The result was a unique (but not very pretty) print on clothes that AI cameras can't see.
To do this, the experts used the SOCO dataset on which the computer vision algorithm YOLOv2 is trained, identified a pattern that helps to recognize a person, created an opposite pattern and transformed it into an image - a print on a sweater. As a result, the owner of such a sweater can hide from detection systems.
"This stylish sweater is a great way to stay warm this winter. It features a waterproof microfleece lining, a modern cut, and anti AI patterns - which will help hide from object detectors. In our demonstration, the YOLOv2 detector was able to fool the detector with a pattern trained on a COCO data set with a carefully constructed target," the researchers note.
Next, they want to release a whole series of clothes with such a print.