Scientists have long been searching for ways to make computer vision more closely resemble human vision, addressing issues such as high energy and memory consumption. These limitations have hindered the effectiveness of technology in critical areas like tracking hypersonic missiles and autonomous navigation.
An innovative approach has been proposed by a team of engineers at the University of Pittsburgh, based on neuromorphic engineering principles inspired by human vision. The researchers aim to develop systems capable of visual information processing with significantly lower energy and data costs, with progress already underway.
Rajkumar Kubendran, an assistant professor at the University of Pittsburgh’s Swanson School of Engineering, has been awarded a $550,000 grant from the US National Science Foundation for developing neuromorphic systems that are highly efficient in terms of energy and data handling.
Kubendran plans to utilize the funding to create a new generation computer vision architecture utilizing biobilized sensors, processors, and algorithms. This will help reduce energy consumption and data transmission volumes, optimizing the technology for operation under resource-constrained conditions.
By applying methods and principles that mimic processes in the human brain’s retina and visual cortex, the approach to visual information processing and perception can be drastically transformed.
The project holds the potential to revolutionize sectors such as healthcare, defense, Internet of Things, and industrial automation, opening up new possibilities for electronics and facilitating the adoption of self-driving vehicles, intelligent video surveillance, and virtual/augmented reality.
Kubendran highlights that while European and Asian companies are currently leading in the development of biobilized visual sensors, the US is lagging behind in this field. His project aims to bridge this gap and secure a dominant position in the deployment of next-generation computer vision systems.