Umut Güçlü at Radboud University in the Netherlands claims these are the most accurate reconstructions to date. The AI system's ability to focus on specific brain regions has been instrumental in improving the quality of the reconstructed images. The study also demonstrated the difference in image quality when the AI system used an attention mechanism versus when it did not.
Key takeaways:
- AI systems can now create accurate reconstructions of what someone is looking at based on recordings of their brain activity.
- The accuracy of these reconstructed images is greatly improved when the AI learns which parts of the brain to pay attention to.
- The reconstructions are said to be the closest and most accurate to date, according to Umut Güçlü at Radboud University in the Netherlands.
- The AI system was tested on a macaque, with the top row of images being the original, the second row being images reconstructed by the AI based on brain recordings, and the bottom row being images reconstructed by the AI system without an attention mechanism.