A breakthrough in the field of robotic vision has been pioneered by a team of computer scientists from the University of Maryland. Building upon the principles of the human visual perception, they have created a camera that not only enhances how the robots perceive their surrounding environment but also how they respond to it. This novel camera system is closely modelled on the involuntary micro-movements occurring in the human visual system to maintain focused and clear visuals over time.
The newly invented system, known as the Artificial Microsaccade-Enhanced Event Camera (AMI-EV), has been explored in detail in a paper published in Science Robotics in May 2024. The study accentuates the importance of enhancing the current event capturing camera technologies to extract blur-free and sharp images during rapid movements.
"Event cameras attempt to track moving objects more efficiently than their traditional counterparts. However, one of their major shortcomings is their inability to produce a sharp image when a fast-paced motion is involved. This issue often poses inevitable problems for autonomous cars and robots, which greatly rely on accurate and timely visuals for adequate response to their changing surroundings," says Botao He, the lead author of the paper.
In quests to replicate human vision abilities, the team was intrigued by microsaccades, which are quick, involuntary eye movements that allow humans to maintain a stable view. Using this concept, the team managed to implement a similar mechanism in the AMI-EV. This was done by inserting a rotating prism inside the AMI-EV to redirect light beams, thereby allowing the camera to stabilize the texture of a recorded object, akin to how a human does.
Moreover, the team's accomplishment is seen as significant in the realm of robotic vision. Co-author Yiannis Aloimonos believes that the invention is a leap forward, saying: "Better cameras mean better perception and reactions for robots."
The potential impacts of the development are thought to extend beyond robotics. The unique features of AMI-EV could be a key solution to the long-standing issues that professionals in industries relying on accurate image capture and shape detection encounter. "Event sensors and AMI-EV are set to revolutionize the field of smart wearables due to their superior performance in extreme lighting conditions, low latency, and low power consumption," says Cornelia Fermüller, senior author of the paper.
In initial tests, AMI-EV was able to accurately capture movement in various contexts. Aloimonos also explains, "Our camera system can solve specific issues, like helping autonomous cars differentiate between a person and other objects on the road. It has numerous potential applications in technologies that the general public interacts with daily," confirming the system's potential to become indispensable technology.
Disclaimer: The above article was written with the assistance of AI. The original sources can be found on ScienceDaily.
