Using artificial intelligence, scientists can reconstruct the images people are thinking about with up to 76 percent accuracy. After measuring brain waves, algorithms can guess the topic of the thought. The results are much better than previous systems, but sometimes still difficult to recognize.
Japanese scientists are using a new way to use artificial intelligence to reconstruct images from human brain waves. They use a so-called Bayesian estimation framework for this purpose, which can successfully reconstruct the images people have seen and the images they have thought about. This is done by measuring brain waves with sensors. The accuracy was 75.6 percent. But it varies greatly from case to case whether you can correctly identify what it is.
Test subjects first viewed 1,200 different images while their brain waves were measured. They were then shown a different picture from the one taken by the 1,200. After half an hour to an hour, they were asked to think about the picture they had seen. A generative AI tool then reconstructed the image based on brain activity. Results It was published this week In the scientific journal Neural Networks.
More complex images
Researchers from the Japan Institute of Quantum Science and Technology (QST) He says This works with all types of images, including landscapes and complex shapes. Until now, reproducing images in this way based on brain activity has only been possible if the person has actually seen them, or if they are fairly easy images, such as faces, letters or simple shapes.
Previous methods were about 50 percent accurate, which the Japanese say represents a major step forward. Researcher Kei Majima speaks in a Japanese newspaper about a “huge achievement.”
Developments can also lead to new forms of communication that do not require words and only ideas are sufficient. Researchers also see applications in reconstructing “phenomena such as delusions, hallucinations, and dreams.”
Read more news about science.
Don’t miss anything, follow Our channel on WhatsApp And receive our newsletter.
“Travel enthusiast. Alcohol lover. Friendly entrepreneur. Coffeeaholic. Award-winning writer.”