265 Bali Medical Journal 2022; 11(1): 265-271 | doi: 10.15562/bmj.v11i1.3096 ORIGINAL ARTICLE ABSTRACT Analysis of RGB range value on fingernail image for detecting diabetes mellitus risk Ima Kurniastuti 1* , Ary Andini 2 , Sabrina Ifahdini Soraya 3 Introduction: Fingernail has various colors related to the organ’s body condition, such as the pancreas, indicated by diabetes mellitus. The study aims to determine and compare the RGB range value on the fingernail image to detect diabetes mellitus risk in fasting and non-fasting conditions. Methods: The study was a true experimental study using fasting and non-fasting respondents. Data were obtained by blood glucose level testing and fingernail image capturing. The result of blood glucose levels was classified into normal, prediabetes, or diabetes, and fingernail images were followed according to their categories. The histogram determined the RGB values of fingernail images, and calculated the maximum value of color intensity based on the height peak appeared. The distribution frequency of each group was used to get a range of RGB fingernail images in each category. Results: Based on the results, it showed a comparison of RGB range value between fasting and non-fasting condition, including range value differences in red and blue, but any slightly overlapped in green range value. In a future study, we will use ordinal logistic regression to determine the prediction program of diabetes mellitus risk. Furthermore, we will develop a program by adding some features to improve the analysis system of the fingernail image for diabetes mellitus risk detection. Conclusion: There was a comparison of RGB range value on fingernail image between fasting and non-fasting condition. Keywords: diabetes mellitus, RGB, Fingernail image, blood glucose, color feature. Cite This Article: Kurniastuti, I., Andini, A., Soraya, S.I. 2022. Analysis of RGB range value on fingernail image for detecting diabetes mellitus risk. Bali Medical Journal 11(1): 265-271. DOI: 10.15562/bmj.v11i1.3096 1 Economy Business and Digital Technology Department, Universitas Nahdlatul Ulama Surabaya, Surabaya, Indonesia; 2 Health Department, Universitas Nahdlatul Ulama Surabaya, Surabaya, Indonesia; 3 National Chiao Tung University, Hsinchu, Taiwan; *Corresponding author: Ima Kurniastuti; Economy Business and Digital Technology Department, Universitas Nahdlatul Ulama Surabaya, Surabaya, Indonesia; ima.kurniastuti@unusa.ac.id Received: 2022-01-11 Accepted: 2022-04-02 Published: 2022-04-17 265 Bali Medical Journal (Bali MedJ) 2022, Volume 11, Number 1: 265-271 P-ISSN.2089-1180, E-ISSN: 2302-2914 Open access: www.balimedicaljournal.org INTRODUCTION e nail could be used as a healthcare diagnosis to detect disease by identifying its color. When fingernail has dominant pink color defined as a “healthy”, but another color defined as a “sick”. e faded color of the nail could be identified as anemia, heart failure, malnutrition, and liver disease. White color on the nail with dark edges is a sign of liver disease such as hepatitis. e yellow color on the nail indicates a fungus infection, thyroid disease, lung disease, diabetes, or psoriasis. While blue color on the nail indicates pneumonia or heart abnormalities. 1 Any change in the color of nails could not be detected using human eyes due to insensitivity to realize small changes in characteristics such as intensity, color, or texture of the object. erefore, an efficient and effective image segmentation method is needed to gather information from images. e image segmentation is initial and most important in image analysis to gather necessary information. Image pixels are grouped according to any characteristic of the image. 2,3 Image analysis will be focused on gathering information according to color information. Image analysis will be focused on gathering information according to color information. As for color information, color space is a mathematical model to represent color information as three or four different color components. 4 One of the color spaces was RGB color space. ree- dimensional Cartesian coordinate systems represent RGB color space by three values red, green, and blue. RGB color space representation is addictive in nature. 5 RGB color space is selected because the simplicity of image-based feature extraction is adequately acceptable, low complexity in computation, and effective in characterizing the distribution of color in an image. 6 e study has focused on nail color information that could be used to detect diabetes mellitus risk. Hyperglycemia is a typical sign of diabetes mellitus due to blood glucose exceeding the normal range (over 200 mg/dl). World Health Organization (WHO) predicts an increase in the number of people with diabetes mellitus, one of the global health threats. WHO predicts an increase in diabetes mellitus patients in Indonesia from 8.4 million in 2000 to around 21.3 million in 2030. is report shows that the number of people with diabetes mellitus is 2-3 times in 2035. Whereas the International Diabetes Federation (IDF) predicts an increase in the number of people with diabetes mellitus in Indonesia from 9.1 million in 2014 it became 14.1 million in 2035. 7 Some researchers have studied early detection of diabetes mellitus by using image color such as iris structure for predicting human health, which is mentioned as iridology. Dr. Bernard Jensen’s Chart of Iris is shown that pancreas activity was shown on Iris in right eyes at direction 07.15 – 07.45 hour. 8,9 Pancreas abnormalities could be indicated by diabetes mellitus. But, the other researcher used energy rates of fingernails image to predict Pancreas condition. Pancreas condition could be predicted in the third step on the right fingernails and the sixth