Try now

Micro-Trend

Algorithm protects from facial recognition

Researchers from the University of Toronto have succeeded in using an algorithm to protect photos from facial recognition technologies. The solution leverages a deep learning technique that pits two algorithms against each other. The first neural network identifies faces, while the other one works to disrupt the facial recognition task of the first, and they both learn from each other. The result is that the algorithm that protects from facial recognition changes specific pixels in the photo that cannot be seen by the human eye but disable the facial recognition algorithm.

Export