Gwangju Institute of Science and Technology Researchers Develop Cat's Eye-Inspired Vision System for Autonomous Robotics

Feline-inspired vision technology enhances accuracy in challenging environments, paving the way for smarter, more efficient autonomous systems

GWANGJU, South Korea, Oct. 16, 2024 /PRNewswire/ -- Researchers led by Professor Young Min Song from the Gwangju Institute of Science and Technology (GIST) have unveiled a vision system inspired by feline eyes to enhance object detection in various lighting conditions. Featuring a unique shape and reflective surface, the system reduces glare in bright environments and boosts sensitivity in low-light scenarios. By filtering unnecessary details, this technology significantly improves the performance of single-lens cameras, representing a notable advancement in robotic vision capabilities.

Autonomous systems like drones, self-driving cars, and robots are becoming more common in our daily lives. However, they often struggle to "see" well in different environments—like bright sunlight, low light, or when objects blend into complex backgrounds. Interestingly, nature may already have the solution to this problem.

Cats are known for their amazing vision in both bright light and darkness. Their eyes are uniquely adapted: during the day, vertical slit-shaped pupils help them focus and reduce glare. At night, their pupils widen to let in more light, and a reflective layer called the tapetum lucidum boosts their night vision, giving their eyes that familiar glow.

A group of Korean researchers led by Professor Young Min Song from Gwangju Institute of Science and Technology (GIST) designed a new vision system that uses an advanced lens and sensors inspired by feline eyes. The system includes a slit-like aperture that, like a cat's vertical pupil, helps filter unnecessary light and focus on key objects. It also uses a special reflective layer similar to the one found in cat eyes that improves visibility in low-light conditions. This research was published in the journal Science Advances on September 18, 2024, and represents a significant advancement in artificial vision systems, demonstrating enhanced object detection and recognition capabilities and positioning it at the forefront of technological breakthroughs in autonomous robotics.

"Robotic cameras often struggle to spot objects in busy or camouflaged backgrounds, especially when lighting conditions change. Our design solves this by letting robots blur out unnecessary details and focus on important objects," explains Prof. Song. This approach has the additional benefit of being energy-efficient, as it relies on the design of the lens rather than on heavy computer processing.

This groundbreaking technology unlocks exciting possibilities for real-world applications, transforming the landscape of robotic vision. The advanced vision system promises to elevate the precision of drones, security robots, and self-driving vehicles, enabling them to adeptly navigate intricate environments and execute tasks with unparalleled accuracy. "From search-and-rescue operations to industrial monitoring, these cutting-edge robotic eyes stand ready to complement or even replace human efforts in a variety of critical scenarios," emphasizes Prof. Song.

Reference        
Title of original paper: Feline eye–inspired artificial vision for enhanced camouflage breaking under diverse light conditions

Journal: Science Advances

DOI: https://doi.org/10.1126/sciadv.adp2809

About the Gwangju Institute of Science and Technology 
http://www.gist.ac.kr/

Contact:
Chang-Sung Kang
82 62 715 6253
[email protected]

SOURCE Gwangju Institute of Science and Technology (GIST)