Computer Vision for Jaundice Detection in Neonates Using Graphic User Interface
Warqaa Hashim, Ali Al‐Naji, Izzat Al-Rayahi, Munir Oudah
Abstract
Title Physiological jaundice occurs in the first week of life in newborns due to the increase in bilirubin level which in turn leads to yellowish discolouration of skin and sclera. Sever jaundice and toxic level of bilirubin can cause brain damage as bilirubin exists in the central nervous systems. Invasive blood sampling is the optimum method to measure bilirubin level; however, it is painful and stressful for the neonate, and it may cause blood loss and can lead to anaemia, especially when repeated blood tests are required. In addition, blood tests expose the infant to the risk of infections. Moreover, invasive tests are time-consuming as their results are not immediate. Due to all the problems mentioned earlier, this paper proposes a new system for jaundice detection that is based on skin colour analysis. The proposed system uses a digital camera as a colour based screening tool as it is affordable, objective, ubiquitous, and less painful to infants. Based on the analysis obtained from the captured images, jaundice was detected and estimated, opening the door for further case studies in medical applications, especially in diagnosis, monitoring patient’s health, and supplying active treatment.