www.zejournal.mobi
December 23, 2024

The futuristic makeup avoids facial recognition systems

Author : Ulyces | Editor : Samus | July 28, 2017 at 02:49 PM

Grigory Bakunov is not just anyone. Director of Technology at Yandex , the equivalent of Google for the Russian-speaking public, he is also the host of a podcast that is listened to each week by more than half a million people. And if there is one thing that frightens Grigory Bakunov, it is the frightening progress of artificial intelligence in surveillance. That is why the Russian engineer developed an algorithm capable of deceiving the most sophisticated facial recognition systems ... with makeup.

Facial recognition is a reality. For example, in Nice, prior to Euro 2016, the technology was tested on street cameras to identify potential trouble-makers. Naturally, it is also the case in Paris, London, New York or Moscow ... in short, all the surveillance systems of large cities are now boosted at the AI. This is not a scoop, even BFMTV is aware. "Facial recognition systems are used by different people for different purposes, and it is impossible to travel to Moscow by avoiding cameras," Grigory Bakunov wrote on the telegra.ph  website  on 14 July. But he found a fault in the matrix.


Read More :

The Russian engineer developed an algorithm capable of hiding the identity of a person in the eyes of artificial intelligence. "A simple but effective algorithm that I developed very quickly," he says. "He was able to generate a facial makeup that could deceive the surveillance cameras, using only a few lines on the face."

If Grigory Bakunov writes to the past, it is because he nevertheless decided not to market the project after realizing that it would offer a simple way for criminals to trick the banks and the police. "That's why I made the decision not to launch it on the market, there was too much risk for it being used for bad purposes.”  He will probably have inspired a few in passing.


- Source : Ulyces

Send via email :

Comment

Send your comment via :



Close

Search
Like Our Site?
(34)
Latest Articles
Most Read Articles
Loading...
Loading...
Loading...

Email Subscribe

Received our newsletter, we send it to your email

  


Close