A team of researchers from the National University of Singapore, the Chinese University of Hong Kong, and Stanford University has used brain scans of research participants to tell what a person is picturing. First, they used fMRIs to record brain signals of participants as they looked at pictures. Researchers then trained an AI model to match brain patterns with specific images. When participants were shown new images, the system detected patient brain waves, generated a description, and created an image based on that description.
The resulting images were similar to many AI-generated images: slightly odd and dreamlike. The idea behind the study is that disabled patients may be able to recover what they see and think, or could lead to humans communicating using just thoughts. There are worries about how invasive this could be, but this research could open many doors in the future.
Read more at NBC.
Comentários