Citation

Karim, Razuan and Zaman, Mukter and Wong, Hin Yong (2022) Implementation Of Smart System for Neonatal Jaundice Detection/Monitoring Using Image Processing. In: Postgraduate Colloquium December 2022, 1-15 December 2022, Multimedia University, Malaysia. (Unpublished)

Text razuan foe.pdf - Submitted Version Restricted to Repository staff only Download (306kB)

Abstract

Hyperbilirubinemia, a condition in which there is too much bilirubin in the blood, is what causes neonatal jaundice and the associated yellow skin color. Raised bilirubin levels can cause brain damage if neglected. At the moment, methods for detecting jaundice require a blood sample, ocular judgment is unreliable, and more precise methodology is expensive. In this study, jaundice was detected and classified using a deep learning-based Convolutional Neural Network (CNN) model with precise tuning. The aforementioned deep neural networks were trained using 800 baby photos, both normal and jaundiced, from a dataset (found on Google). A second dataset of 100 newborn photographs of Bangladeshi babies was obtained and utilized for testing and prediction. Numerous important metrics, including precision, F1-score, recall, specificity, confusion matrices and loss graphs were computed to verify the model's accuracy. F1-score, recall, precision and specificity accuracy achieved were 0.99%, 0.99%, 100%, and 100%, respectively. The model was then tested using a dataset, and the evaluation findings revealed that the suggested CNN-based model had a 99.38% accuracy rate.

Item Type:	Conference or Workshop Item (Poster)
Uncontrolled Keywords:	Image Processing, bilirubin
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) > TA1501-1820 Applied optics. Photonics
Divisions:	Faculty of Engineering (FOE)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	28 Dec 2022 01:11
Last Modified:	28 Dec 2022 01:11
URII:	http://shdl.mmu.edu.my/id/eprint/11017

Downloads

Edit (login required)