RMIT University researchers have developed an early-stage “neuromorphic vision” prototype designed to combine sensing, memory and information processing in a single system, a step they say could eventually support smart bionic eyes and other low-power machine vision applications.
The university said the prototype aims to reduce the amount of data transferred between separate sensors, memory and processors by performing more processing at the point where visual information is captured. RMIT described the approach as potentially reducing both data volume and energy use for complex visual tasks, though it remains a research demonstration.
RMIT has filed an international patent application under the Patent Cooperation Treaty for the invention. The work is led by Professor Sumeet Walia at RMIT’s Centre for Opto-electronic Materials and Sensors (COMAS) and builds on research involving atom-thin semiconductor materials, including molybdenum disulfide (MoS₂).
“Nature has already solved many of the challenges we’re trying to address in electronics,” Walia said.
“The human eye and brain work together incredibly efficiently, processing vast amounts of information using remarkably little energy. Our research is helping lay the foundations for technologies that work in a more similar way.”
RMIT said the prototype differs from conventional camera systems by aiming to process information locally rather than sending large volumes of data to separate computing components. The university said sensing, processing and storage take place on a 2cm by 2cm chip, housed within a larger 15cm by 14cm by 3cm prototype that includes electronics to read, process and communicate information.
According to RMIT, the system has been trained to recognise patterns such as numbers, shapes and movement, and in laboratory tests can detect changes in what it sees, store that information as memory, and process it locally.
Dr Taimur Ahmed, a co-researcher at RMIT working on neuromorphic vision devices, said the goal was to more closely mirror biological vision by integrating sensing and processing.
“This is not just a sensor that captures information, it’s a sensor that can also process information,” Ahmed said.
“Rather than constantly moving data between separate memory and processing units, much of that work happens much closer to where the information is generated.”
The university said a smart bionic eye would need to identify significant changes in a scene, store relevant information and process visual signals quickly while consuming minimal energy, and argued its approach could help by filtering and interpreting visual information at the point of sensing.
RMIT also highlighted work on a water-based fabrication process intended to transfer atom-thin semiconductors and electrodes with fewer defects than conventional methods, which it said improved electrical and light-sensing performance.
“Each breakthrough brings us closer to technologies such as a smart bionic eye,” Walia said.
While the researchers said practical applications remain years away, RMIT suggested the technology could also support advanced machine vision systems, autonomous vehicles, robotics and intelligent sensors. The university positioned the work as relevant to efforts to reduce the energy demands of artificial intelligence by processing data closer to where it is collected, rather than relying on large-scale data centre processing.
“This work combines advanced materials, engineering and artificial intelligence to address one of the defining challenges of our time: creating intelligent systems that are both powerful and sustainable,” Walia said.
“If this RMIT technology can be scaled up, it could help reduce the amount of data that needs to be moved, stored and processed, making future AI systems more energy efficient.”
RMIT cited two related studies: “PVA-mediated transfer of MoS₂ and Au electrodes: a lithography-free route to ultraclean van der Waals interfaces for high-performance electronics and optoelectronics”, published in ACS Applied Materials and Interfaces (DOI: 10.1021/acsami.6c05536), and “Photoactive monolayer MoS₂ for spiking neural networks enabled machine vision applications”, published in Advanced Materials Technologies (DOI: 10.1002/admt.202401677).
