A study by Australian scientists has shown that even if we believe the images we see are real, our brains are adept at recognizing complex deep frauds. It turns out that the responses to images of people present and those generated by a realistic neural network are different – and this is read using an electroencephalogram (EEG).

Details of the study

The results were obtained during tests designed to assess the human response to deceptive events. alsoImages including images of obviously fake people, cars, interiors. Interestingly, this even works for inverted images.

Understood that even with a cursory glance, observers’ brains can recognize fake faces. However, people found it more difficult to distinguish a real person from a fake, and in some cases viewed fake pictures as more real than real ones.

Although the brain can tell the difference between real and realistic individuals, observers cannot consciously tell them apart. Our findings regarding the divergence between brain response and behavior have implications for how we examine false face perception.
– tell the authors of the work.

In their study, the researchers used images created by Generative Adversarial Networks (GAN) shared by NVIDIA. The dataset included 25 people, cars and bedrooms with processing levels ranging from “unrealistic” to “realistic”.