IIT Guwahati Creates Brain-Inspired Device for Low-Energy Visual Processing
Researchers at IIT Guwahati develop an ultrathin artificial synapse for low-energy visual processing and facial recognition.
In a major breakthrough for electronic image processing, researchers at the Indian Institute of Technology Guwahati, in collaboration with IIT Kharagpur, have developed a light-sensitive device that mimics the human brain's ability to sense, store, and process visual information with minimal energy consumption.
Mimicking the Human Brain
Traditional electronic image processing relies on separate components—a camera for capture, memory for storage, and a processor for analysis. This separation requires massive data transfer, leading to high energy expenditure. In contrast, the human brain performs these tasks simultaneously using a vast network of connections called synapses.
The research team has successfully developed an ultrathin artificial synapse that can respond to light and store information even after the light source is removed. This "memory-like" effect allows the device to function much like the human brain, which recognizes objects instantly by combining sensory input with stored context.
The Science Behind the Breakthrough
To achieve this neuromorphic capability, the team utilized two key materials:
- MXene nanosheets: An electrically conductive, ultrathin 2D material.
- P3HT: An organic semiconductor capable of producing electrical charges in response to light.
When these materials are brought into contact, light-induced electrical charges separate at the interface and remain separated for an extended period. This creates a device that combines sensing, memory, and processing into a single platform.
High Efficiency and Learning Capabilities
Professor PK Giri, the lead author and Professor in the Department of Physics at IIT Guwahati, highlighted the device's ability to "learn." Much like human neurons strengthen through repeated stimulation, the device shows an increased response with repeated signals. In tests, a second light signal elicited a response 180% stronger than the first.
Key performance metrics from lab tests include:
- Energy Efficiency: Consumes 500 times less energy than conventional MXene-based devices.
- Facial Recognition: Achieved 90% accuracy in testing.
- Traffic Scene Segmentation: Achieved 85% accuracy.
Future Applications
This low-cost, solution-processed technology is poised to revolutionize several sectors, including:
- Intelligent facial recognition systems.
- Artificial and robotic vision.
- Smart computing applications.
The research findings have been published in the prestigious journal Carbon. The team included researchers from both IIT Guwahati and IIT Kharagpur, contributing to a significant leap in neuromorphic engineering.