We have reached that period in the history of science and technology where the most recent advances seem inspired by relatively recent works of fiction, such as Harry Potter.
At least that seems to be the story behind it. InvisDefense, the prototype of an invisibility cloak developed by scientists at Wuhan University in China.
However, this cloak has much more modest powers. Since it does not actually make its users invisible to the human eye.
But it has the enviable quality of completely hiding the wearer from Chinese surveillance cameras powered by Artificial Intelligence systems.
The issue of observation and the use of biometrics to identify citizens in a regime like that of that country has been a topic of debate for years.
So the development of this type of layer represents a solution for those who are still looking for a way to maintain a bit of privacy in a country where the advanced use of these surveillance mechanisms is already commonplace.
This is how the InvisDefense invisibility cloak works:
the colleagues of VICE WORLD NEWS have published an extensive article where they report the details of the conception and development of the InvisDefense system, by a team of researchers from the University of Wuhan.
Where they went through more than 700 failed prototype attempts over three months. Where each test used the University campus’s own security cameras and their biometric detection functions.
This is how Wei Hui, responsible for the algorithm that has made the operation of the shelter possible, relates the evolution of the project:
“The hardest part is balancing the camouflage pattern; traditionally researchers used bright images to interfere with machine vision, and it worked, but that makes it conspicuous to human eyes, making the user even more visible.
The fact that security cameras cannot detect the InvisDefense coat means they are flawed and machine vision researchers could use our algorithms to improve current models.”
In the end, the scientists found that the accuracy of pedestrian detection was reduced by 57% after refining the pattern of patches and their arrangement to confuse the algorithm of the thermal cameras.
This layer, according to its own creators, was developed to protect privacy or unwanted detection of people. It’s not perfect, but it’s an important advance.