This study aims a mobile support system to aid health care profesionals in hospitals or in regions far away from hospitals to utilize noninvasive image processing methods for classification of neonatal jaundice. A considerably low processing cost is aimed to be attained by developing an algorithm that could work on a mobile device with low-end camera and processor capabilities within this study. In this context, an algorithm with low cost is developed performing detection of most meaningful parameters by a multiple input single output regression model and correlation.The advantage of the proposed method is that it can estimate bilirubine with the help of a simple regression curve. The reason for its low cost is that the non-invasive jaundice prediction is performed with a simple regression curve instead of many mathematical operations in morphological image processing methods. The study was performed on a total of 196 subjects, 61 among which were classified as severe jaundice while 95 of the newborns were mild jaundice cases and other 40 cases are used for tests. As a result of this work, the two-group classification accuracy of the developed algorithm is observed to be 92.5% for the 40 subject test group.

Cardiotocography (CTG), which measures the fetal heart rate (FHR) and maternal uterine contractions (UC) simultaneously, is used for monitoring fetal well-being during delivery or antenatally at the third trimester. Baseline FHR and its response to uterine contractions can be used to diagnose fetal distress, which may necessitate therapeutic intervention. In this study, a machine learning model based on feature extraction (autoencoder), feature selection (recursive feature elimination), and Bayesian optimization, was proposed to diagnose and classify the different conditions of fetuses (Normal, Suspect, Pathologic) along with the CTG morphological patterns. The model was evaluated on a publicly available CTG dataset. This research also addressed the imbalance nature of the CTG dataset. The proposed model has a potential application as a decision support tool to manage pregnancies. The proposed model resulted in good performance analysis metrics. Using this model with Random Forest resulted in a model accuracy of 96.62% for fetal status classification and 94.96% for CTG morphological pattern classification. In rational terms, the model was able to accurately predict 98% Suspect cases and 98.6% Pathologic cases in the dataset. The combination of predicting and classifying fetal status as well as the CTG morphological patterns shows potential in monitoring high-risk pregnancies.

No abstract