Study: Unnoticed eye movements can be key for better self-driving cars

Jun 23, 2022

New Delhi, June 23 (ANI): An artificial neural network that learns to recognise objects faster and more accurately was developed by Andrea Benucci and colleagues at the RIKEN Brain Science Institute. This study, recently published in the journal PLOS Computational Biology, focuses on all the overlooked eye movements we do. By enabling stable recognition of objects, we show that they serve an important purpose. These findings can be transferred to machine vision, for example, to make it easier for self-driving cars to learn important features on the road. Despite constant head and eye movements throughout the day, the constant changes in physical information that hits the retina do not blur or obscure objects around the world. It is the neural copy of motor commands that is likely to enable this perceptual stability. These copies are sent through the brain each time we move and are designed to allow the brain to adapt to our own movements and stabilise our perception.