Electronic fetal monitoring provides useful information during labor and it allows obstetricians to monitor the status and health of the fetus. In spite of all the current technological advances, there is often poor inter and intra reliability of Fetal heart Rate (FHR) pattern interpretation among obstetricians. The research will examine the hypothesis """"""""Can an artificial neural network interpret FHR data at the level of human experts"""""""".
The specific aims and objects of this research will be 1. To examine the task of aiding the physician in FHR interpretation by applying feedforward artificial neural networks (ANNs) methods to analyze and classify FHR data. 2. To study the use of ANNs to model and enhance key characteristics of the FHR signal and to correlate and classify the data to detect fetal abnormalities and dysfunctions. 3. To investigate and incorporate computer methods to aa) display and explain FHR data and ANN activity, (using scientific visualization methods), b) derive the key mechanisms for the ANN's conclusions, and c) determine the confidence levels that the ANN has in its conclusions. Gains from this research should permit the real time processing, analysis, and interpretations of FHR signals, and the real time classification of the state of the fetus to determine if early intervention is necessary.
Showing the most recent 10 out of 99 publications
