A Computer Can Now Predict Your Thoughts And Draw Images Based On Them

A team of researchers at the University of Helsinki has developed a first of its kind technique in which a computer models your visual perception by monitoring brain signals.

Basically it is like computer trying to imagine what you are thinking about and thus the technique have resulted in completely fictional images.

The research recently appeared in the Scientific Reports journal.

This new technique is basically based on a brain-computer interface (BCI). BCI is not a new thing but most of the previous devices just had one way communication wherein signals are transferred from brain to computer only.

But in this case, both the computer’s presentation of the information and brain signals were modeled simultaneously using artificial intelligence methods. Images that matched the visual characteristics that participants were focusing on were generated through interaction between human brain responses and a generative neural network.

With this modeling called neuroadaptive generative modeling, the team evaluated the effectiveness of the technique with the help of 31 volunteers who participated in the study. These people were shown hundreds of AI-generated images of diverse-looking people and were told to focus on specific features like faces that looked old or were smiling.

While the volunteers were focusing on the features, EEGs of their brain were fed to a neural network which then generated images based on that data. The volunteers then evaluated the computer-generated images and they matched nearly perfectly. The outcomes were really good with an accuracy of 83%.

“The technique combines natural human responses with the computer’s ability to create new information. In the experiment, the participants were only asked to look at the computer-generated images. The computer, in turn, modelled the images displayed and the human reaction toward the images by using human brain responses. From this, the computer can create an entirely new image that matches the user’s intention,” says Tuukka Ruotsalo, Academy of Finland Research Fellow at the University of Helsinki, Finland and Associate Professor at the University of Copenhagen, Denmark.

Journal Reference:
Lauri Kangassalo, Michiel Spapé, Tuukka Ruotsalo. Neuroadaptive modelling for generating images matching perceptual categories. Scientific Reports, 2020; 10 (1) DOI: 10.1038/s41598-020-71287-1

Press Release: University of Helsinki