A Fast and Effective System for Detection of Neonatal Jaundice with a Dynamic Threshold White Balance Algorithm

Abstract

Neonatal jaundice is caused by high levels of bilirubin in the body, which most commonly appears within three days of birth among newborns. Neonatal jaundice detection systems can take pictures in different places and upload them to the system for judgment. However, the white balance problem of the images is often encountered in these detection systems. The color shift images induced by different light haloes will result in the system causing errors in judging the images. The true color of images is very important information when the detection system judges the jaundice value. At present, most systems adopt specific assumption methods and rely on color charts to adjust images. In this study, we propose a novel white balance method with dynamic threshold to screen appropriate feature factors at different color temperatures iteratively and make the adjustment results of different images close to the same. The experimental results indicate that the proposed method achieves superior results in comparison with several traditional approaches.

A Non-invasive Methods for Neonatal Jaundice Detection and Monitoring to Assess Bilirubin Level: A Review

“…Diverse sections of a neonate's body, including the face [34], forehead [2, 17-18, 21, 35], sternum [5-6, 18, 35], abdomen [5,36], or other body parts such as the arm, palm and sole [2] and face, feet, arms and central body [37] have been utilized for skin-based conclusions. The authors adopted various methodologies for highlight extraction, including mean, skewness, standard deviation, energy, kurtosis, and entropy [34], YCbCr and lab color spaces [6,[35][36], RGB [2,6,21,36,37], hue and saturation values [5,18], and diffuse reflection spectral characteristics [2]. Diverse machine learning models, such as KNN [36,38], SVR/SVM [2,36], regression [5,21,37], and an ensemble of classifiers, counting KNN, LARS-Lasso elastic net, LARS, SVR, and RF [6,35], have been used to determine jaundice.…”

“…The authors adopted various methodologies for highlight extraction, including mean, skewness, standard deviation, energy, kurtosis, and entropy [34], YCbCr and lab color spaces [6,[35][36], RGB [2,6,21,36,37], hue and saturation values [5,18], and diffuse reflection spectral characteristics [2]. Diverse machine learning models, such as KNN [36,38], SVR/SVM [2,36], regression [5,21,37], and an ensemble of classifiers, counting KNN, LARS-Lasso elastic net, LARS, SVR, and RF [6,35], have been used to determine jaundice.…”

“…The items in [5-6, 21, 35, 40], but not [34,38], utilized a calibration card for color adjustment in advance. In [35], the authors developed a revolutionary white-adjustment technique with an active edge for modifying various color temperatures without a calibration card. The works cited in [18,34] gathered bilirubin coloring in stripes from serum.…”

Neonatal Jaundice Diagnosis Using a Smartphone Camera Based on Eye, Skin, and Fused Features with Transfer Learning

“…Several methods have been adopted in the literature for data collection. A smartphone camera has been used successfully to capture images of jaundiced and healthy neonates, such as in [8,9,11,12,15,16,18,20,22,38,39]. In contrast, the work in [8] tested several methods, including a direct camera method, a yellowish-green gelatin filter method, and a dermatoscope method, to determine whether a smartphone camera can be used as a screening tool for jaundice.…”

“…Further, the studies in [9,[11][12][13][14] used a calibration card for the purpose of color balancing, while the studies in [7,38] did not. In [39], the authors proposed a novel white balancing method with a dynamic threshold for adjusting different color temperatures without the use of a calibration card. The works in [10,38] collected serum bilirubin coloration on stripes.…”