Research on Estimation of Blood Glucose Based on PPG and Deep Neural Networks

Abstract

Diabetes is one of the three major chronic diseases in the world. At present, the number of diabetic patients in the world is increasing year by year, which has become one of the main threats to us. Therefore, it is very important to monitor blood glucose level. Clinically, blood glucose is measured by collecting fingertip blood, but this method has many disadvantages. In this paper, PPG signals are used to estimate BGL using deep neural networks (DNN). Finally, we found that the success rate of our DNN blood glucose estimation model reached 90.25%, and achieved good results. It provides a choice for the commercialization of noninvasive blood glucose detection technology.