Development of a Smartphone-based diagnostic Tool for Jaundice

Abstract

Jaundice is the cause of an estimated 114,100 deaths among newborn annually. Most of these deaths occur in the poorest regions of the world where current diagnostic technologies are too expensive. In this thesis, several methods were tested using an HTC One V and a Samsung Galaxy S3 to see whether smartphones can be used as an affordable diagnostic tool for jaundice. The methods include attaching foldable spectrometers to the phones, bandpass filters, pressing the camera lens onto the skin, and using color analysis of images taken with the phones. All methods were tested qualitatively using my own skin with either a bruise or with carrot juice applied in order to simulate an increased bilirubin concentration. The color analysis of images was also quantitatively compared to numerical simulations of skin using optical diffusion theory.

Color analysis of images taken with the phones was found to be the most promising of all the methods. The standard deviations of the color responses of both cameras were measured to be significantly smaller than the color variation of skin due to increased bilirubin concentrations. A color calibration technique intended to provide accurate enough calibrations of the cameras for the phones to be used for bilirubin concentration measurements was developed. In its current state, it does not produce calibrations of the needed quality. A roadmap is therefore presented for the further development of the technique needed to yield satisfying results.