Ali, Nurashlida (2023) Ann-based non–invasive jaundice measurement system using optical technique. Masters thesis, Universiti Tun Hussein Onn Malaysia.

Preview	Text 24p NURASHLIDA ALI.pdf Download (553kB) | Preview
	Text (Copyright Declaration) NURASHLIDA ALI COPYRIGHT DECLARATION.pdf Restricted to Repository staff only Download (419kB) | Request a copy
	Text (Full Text) NURASHLIDA ALI WATERMARK.pdf Restricted to Registered users only Download (22MB) | Request a copy

Abstract

Jaundice is a common condition in neonates and occurs due to an increase in the concentration of bilirubin in the body. The conventional method of detecting this condition is through an invasive procedure which extracts blood samples from the neonate's vein in order to assess the concentration value of bilirubin. However, this method might be traumatizing for neonates. The research undertaken in this work mainly focused on a non-invasive jaundice detecting system that utilized optic energy to minimize neonates’ discomfort relative to the conventional method. In this system, the blue LED with electromagnetic spectrum wavelength, 470nm was used as a transmitter. The transmitter sent a spectrum of optic energy to the bilirubin, resulting in reflected optic energy which was measured using a photodiode. The measured data was then captured using an microcontroller and transferred to the computer for further classification of ‘normal’ or ‘jaundice’ status using the Artificial Neural Network. To assess the developed system, the optical energy reflection values detected in voltage units were mapped to the bilirubin concentration values. Based on that, it was decided if the condition of the neonate had been classified correctly, whether normal or jaundiced. The jaundice meter prototype developed in this study was tested among 44 patients, and the validation results showed an accuracy of 90% with JM103. This research concluded that the proposed system can be used as an alternative method of assessing bilirubin concentration among neonates

Actions (login required)

View Item