Deep Learning based Colorimetric Classification of Glucose with Au-Ag nanoparticles using Smartphone

Abstract

Glucose is an extremely important molecule as an energy source and human body function. Diabetes, which ranks among the diseases of the age, is detected based on the glucose level in the human body. Therefore, quantification of glucose is important to develop research and applications of diabetes, which is an important health problem. This study aims to classify glucose concentration with deep learning based colorimetric analysis using a smartphone. The color changes obtained as a result of the reaction of Au-Ag nanoparticles with different concentrations of glucose were captured using a smartphone camera to create a dataset. The proposed deep learning model was trained with this dataset and glucose concentration was classified with 95.93% accuracy. The deep learning model was integrated into our custom-designed Android application, DeepGlucose, to enable glucose classification via a smartphone.