Glucose Content Analysis using Image Processing and Machine Learning

Technology is constantly evolving to make it easier for people to work in biomedical research and technology daily. The glucose level checking system can use a urine analyzer detector as a color reader of the urine strip. This work aims to analyse glucose levels based on digital picture identification using the MATLAB application for patient glucose data processing. Injections are joint for diabetic people to control their blood sugar levels. 

Repeated injections might cause minor physical harm to the body that can hamper the immune system's ability to fight against pathogens. Numerous research has concentrated on non-invasive glucose-based testing, namely using urine. This study was created using image processing to examine the non-invasive glucose testing procedure. The noise is cleaned up using a Gaussian filter and histogram-based feature extraction for picture database extraction. 

Support vector machines classify data using a 70% training and 30% testing process. The SVM classification results had an accuracy of 85% and time processing of 0.5 seconds. In making medical decisions, it is possible to consider the effects of diabetes, pre-diabetes, and non-diabetes.

Keywords: glucose, gaussian filter, histogram feature, support vector machine.