Identifying Glucose Levels in Human Urine via Red Green Blue Color Compositions Analysis

Abstract

Diabetes mellitus (DM), a metabolic disorder caused by the lack of the insulin hormone, has become a health problem quite severe and is the most common endocrine disease. Recently, diagnosing diabetes could be carried out through monitoring the glucose level in human blood taken from the patient's finger or arm. On the other hand, a non-invasive blood sugar detector with a benedict test on human urine is an alternative to monitor blood sugar without injuring the body. The test output can be determined from the colour of the colour change of urine. However, manual evaluations on the urine colour are prone to human subjectivity. In this paper, we present a computational method to determine the blood sugar level based on the colour of the given urine automatically. The proposed method identifies the blood sugar level by taking into account the colour intensity on the red, green, and blue (RGB) channels of the urine colour. In the experimental parts, the proposed method is capable of classifying the urine sample correctly. Hence, our approach may be beneficial for practical applications.