AI-Enabled Ultrasound Data Analysis for Reproductive Health Monitoring

Reproductive health surveillance is important in ascertaining fetal and maternal well-being, and ultrasound and cardiotocography (CTG) are the most prevalent non-invasive diagnostic modalities. Nonetheless, standard analysis of these data highly depends on clinician judgment, which may cause subjectivity and delay in decision-making. This project introduces an AI-Enabled Ultrasound Data Analysis System that aims to categorize fetal health status based on categorical ultrasound and CTG parameters. The framework uses machine learning models such as Logistic Regression, Random Forest, and XGBoost on a well-structured fetal health dataset. Of these, the XGBoost model demonstrated better performance with accuracy up to 96%, clearly separating Normal, Suspect, and Pathological cases. A web interface was implemented to support real-time user input, allowing clinicians and medical practitioners to enter diagnostic parameters and receive fetal health predictions instantly. The combination of categorical data analysis with AI improves diagnostic accuracy, minimizes inter-observer variation, and offers a scalable method for reproductive health monitoring. This research illustrates that machine learning-based analysis of ultrasound and CTG characteristics can aid early detection of fetal distress and help healthcare professionals make robust, evidence-based clinical decisions.