Background

Electronic fetal monitoring (EFM) involves simultaneous monitoring of the fetal heart rate (FHR) as well as the maternal uterine pressure (UP) signal. In practice, EFM has been based on visual analysis of FHR patterns and thereby is rather subjective. A reliable computer-based system for accurate and meaningful classification of FHR signals does not exist on the market. The obstetric practice needs a system to fill the void.

Technology

Presented is a system for receiving FHR and UP signals and extracting feature values from these signals. The features are used for online classification exploiting the Bayesian machinery. The system provides a real time output summarizing the state of the fetus which can range from fully healthy to abnormal. With the system, an obstetrician has a reliable decision support system that will benefit the fetus, the mother, and the obstetrician.

Advantages

Sequential real time classification - Assimilation of physician feedback To the system for enhanced learning - data visualization for quantification of classification uncertainty - Encoding of temporal effects - Knowledge Discovery of FHR categories without assistance of gold standard labeling

Application

Classify the fetal data records and assign to the classification a measurement of uncertainty.