Guwahati: Researchers at IIT Guwahati, in collaboration with IIT Kharagpur, have developed a brain-inspired, light-sensitive device that consumes 500 times less energy for visual information processing, potentially improving the energy efficiency of image-processing applications.The device can detect visual information, store it and process it, mimicking the way the human brain responds to and remembers what it sees. The developed technology can be used in facial recognition, artificial and robotic vision, and other smart computing applications, IIT Guwahati said in a media release on Tuesday.Globally, scientists are developing neuromorphic (brain-like) systems that mimic the human brain. A key component of these systems is an artificial synapse, which mimics the connections between neurons. Conventional electronic image processing systems use separate components for image capture, storage and processing, requiring substantial data transfer and energy consumption. The human brain, however, performs these functions with remarkable efficiency through an extensive network of connections between nerve cells, known as synapses.The research was led by Prof PK Giri of IIT Guwahati’s department of physics. He and his research team have developed an ultrathin artificial synapse that responds to light and retains information even after the light source is switched off. This resembles the human brain’s ability to retain visual information even after an object is no longer in view.“When we capture the visual of an apple through electronic vision, the camera captures the light from the apple, converts it into digital information and sends it to a microprocessor to analyse the image pixel by pixel to identify it as an apple. On the other side, when we see an apple with our eyes, the brain immediately recognises it using the already stored information about colours, shapes, and related context. This reduces the need to consciously and sequentially analyse the image to identify any object,” Giri, lead author of the research, said.He said that a brief light signal generates an electrical response that persists even after the light is switched off. Furthermore, repeated light signals can strengthen the response.“In the developed device, a second light signal elicited a 180% stronger response than the first. This is similar to our neurons learning through repeated stimulation. Thus, our device combines sensing, memory, and information-processing functions in a single platform,” Giri said.The findings have been published in the journal Carbon. The research paper was co-authored by Giri and his team members Sanjoy Sur Roy, Sanju Nandi, Debabrata Sahu, Garima Choudhary and Prof M Meyyappan from IIT Guwahati, and Saswata Goswami and Prof Chandra Sekhar Tiwary from IIT Kharagpur.
