jScan: Smartphone-Assisted Bilirubin Quantification and Jaundice Screening

Jaundice occurs due to an imbalance in the levels of bilirubin in the body. Various invasive and noninvasive methods are used to identify the bilirubin level. However, the existing approaches are time-consuming, costly, require additional setup, and some are invasive. To address these issues, this article proposes a smart jaundice diagnosis method. The system utilizes a smartphone camera to capture eye images and automatically extract the sclera region. It performs a color quantization process to determine the bilirubin level from the sclera. If the bilirubin level
${>}{3}$
mg/dL, the person is classified as hyperbilirubinemic. The model achieved
${R}^{{2}} = {0.956}$
with a bias of 0.43 mg/dL and a sensitivity of 92.85% for 102 participants. The model works with different lighting conditions and devices, demonstrating robustness, and independence. To make it available and affordable in remote areas for continuous monitoring, the model is converted into a mobile application.