Bioinspired adaptive pupil reflex based on liquid-metal shape-shifters for machine vision

Area of Science:

• Bioinspired engineering
• Artificial vision systems
• Robotics

Background:

• Biological eyes exhibit remarkable adaptation to diverse environments.
• Existing artificial vision systems face challenges in adapting to drastic environmental changes, particularly light intensity.
• Developing artificial pupils with dynamic adjustment capabilities is crucial for advanced machine vision.

Purpose of the Study:

• To develop a bioinspired vision system integrating an artificial retina, liquid-metal visual neurons, and an adaptive artificial pupil.
• To simulate closed-loop pupil reflex behavior for enhanced visual recognition.
• To demonstrate the system's ability to emulate various animal pupil shapes and functions.

Main Methods:

• Integration of a hemispherical imaging array as an artificial retina.
• Utilization of liquid-metal shape-shifters as programmable visual neurons capable of simulating nerve signals.
• Implementation of an adaptive artificial pupil with controlled liquid-metal deformation for light regulation.

Main Results:

• The system successfully simulated biological spike nerve signals through rapid switching of liquid-metal circuits.
• Adaptive pupil deformation under strong light improved image recognition accuracy in high-light conditions.
• Demonstrated key functionalities including ultrawide field of view, adaptive light adjustment, and image recognition.

Conclusions:

• The proposed bioinspired vision system effectively simulates animal pupil functions and adaptive light responses.
• The system's programmability and ability to mimic diverse pupil shapes show significant potential for robotic systems and advanced machine vision.
• This research offers a pathway towards more robust and adaptable artificial vision technologies for applications like autonomous driving.