A collaborative research team has developed a new method for measuring glucose using deep learning technology, as published in Laser & Photonics Reviews. The team optimized a sensor based on split ring resonators (SRRs) to address issues of inconsistent measurements influenced by factors like temperature and humidity. They used photolithography to create patterns on semiconductors and employed deep learning technology to enable the sensors to learn from electrical signals at various locations. The team developed a one-dimensional convolutional neural network (1D CNN) that effectively compensated for errors due to sample location variations, achieving low mean absolute error and mean squared error. The team included researchers from Pohang University of Science and Technology (POSTECH) and Daegu University, with Professor Junsuk Rho highlighting the enhanced consistency and reliability of the glucose measurement device. The technology can be commercialized and mass-produced using existing photolithography processes. The research was published in Laser & Photonics Reviews, with more information available in the published paper. This advancement in glucose measurement technology shows promise for improving accuracy and reliability in various applications.
Source link
Source link: https://phys.org/news/2024-07-glucose-food-easier-deep.html
#Simplifying glucose measurement in food with deep learning technology. #Healthcare
