Researchers at MIT have developed a way to enhance AI models by simulating peripheral vision, a capability that humans possess but machines lack. By training computer vision models with an image dataset that mimics peripheral vision, the researchers found that the models improved in detecting objects in the visual periphery. However, the AI models still did not perform as well as humans in this aspect. The study also revealed that factors like object size and visual clutter did not strongly impact the AI’s performance, raising questions about what is missing in these models compared to human vision. This research could have implications for improving driver safety and understanding human behavior. The findings were detailed in a paper by MIT researchers, and for more information, readers can access the full article on the MIT News website.
Source link
Source link: https://www.predictiveanalyticsworld.com/machinelearningtimes/researchers-enhance-peripheral-vision-in-ai-models/13438/
AI models’ peripheral vision improved by researchers #innovation
Preeti
1.8k Views
0 Votes
GIPHY App Key not set. Please check settings