Study on Non-Invasive Blood Glucose Monitoring Methods for Japanese Black Cattle

This paper proposes a method for non-invasively measuring blood glucose levels in beef cattle. Proper feeding management is necessary to improve productivity in the livestock industry, and blood tests are effective as a means of monitoring the health status of cattle with feed. However, blood tests are very labor intensive because they require holding cattle and specialized technicians to perform them. A noninvasive method for estimating blood glucose levels in humans has been proposed, using absorbance spectra and the artificial neural network (ANN) obtained by irradiating blood vessels with near-infrared light. In this paper, we will examine whether it is possible to classify blood glucose levels using a convolutional neural network (CNN) based on absorbance spectra of blood vessels obtained with a small near-infrared spectrometer. The proposed method had better blood glucose classification performance than ANN, suggesting its effectiveness as a noninvasive blood glucose measurement method.