Fetal Distress Classification
An LSTM classifier to predict fetal distress based on fetal heart rate and uterine contraction time series data.
Jan 2019 - March 2019
Fetal distress before and during childbirth indicates that the fetus has been receiving inadequate oxygen. The goal was to develop a system that classifies a fetus as “distressed” or “normal” using fetal heart rate (FHR) and uterine contraction (UC) time series obtained from cardiotocography. We used the CTU-CHB Intrapartum Cardiotocography Database which consists of 552 readings sampled at 4 Hz. We extracted prominent features such as accelerations and decelerations from the data, and developed an LSTM classifier. My team and I won the first place in our category at the 2019 Smart India Hackathon for our application to predict fetal distress.