In recent years, artificial intelligence has significantly influenced the medical field, driving advancements from improved cancer screening to accelerated drug discovery. Breakthroughs in this area are emerging at an extraordinary rate.
Now, an innovative app powered by AI is poised to make a difference for parents facing concerns about infant jaundice, a condition that is relatively common. This application, created by Singapore General Hospital (SGH), SingHealth Polyclinics (SHP), and the national HealthTech agency Synapxe, enables parents to screen their newborns for neonatal jaundice (NNJ) effectively.
The app uses a pioneering machine learning model tailored for Asian skin tones, delivering impressive accuracy. According to SingHealth Polyclinics, the app showed identical results compared to traditional NNJ testing methods, achieving a remarkable sensitivity rate of 100%.
Named BiliSG, this app was developed after two years of clinical research involving validated studies with infants under 21 days old. It utilizes machine learning to evaluate skin coloration in areas such as the chest, forehead, and abdomen. 
The name BiliSG is derived from bilirubin, a yellow pigment produced during the breakdown of red blood cells, which is also a vital part of bile generated by the liver. In cases of jaundice, elevated bilirubin levels can cause a yellowish hue to the skin due to its seepage into skin tissues. This app simplifies the process for parents to check for signs of neonatal jaundice in the comfort of their homes, reducing the need for frequent hospital visits.
NNJ affects around 60% of infants, with even higher rates observed in premature babies. While the body usually regulates bilirubin levels, failure to do so can lead to severe complications, such as permanent brain damage, developmental delays, and motor disabilities.
According to medical researchers involved in this initiative, “This app will revolutionize how we safeguard infants’ brain health by allowing parents to monitor jaundice closely in a convenient setting, supported by telemedicine consultations with their family doctors.”
To use the app, parents simply take a photo of the skin area marked by a specially designed color calibration sticker. This sticker minimizes reflections, adjusts for variations in light intensity, and accounts for temperature fluctuations. The development team is currently working to refine the app through ongoing validation efforts.
Initial tests were conducted using an iPhone 12, but the research group is eager to broaden compatibility to include more smartphone models, such as those running on Android. This expansion, however, presents challenges related to camera specifications (like resolution) and different lighting conditions that must be meticulously addressed.
The project’s team is focused on making the BiliSG app more robust and reliable for everyday use beyond clinical environments, with an emphasis on accessibility for parents and caregivers who may lack experience with medical devices.
For more in-depth information, you can access the complete research paper here.
