Artificial Intelligence in Cardiac Topographic Ultrasonographic Screening: The Journey to Gitam SHK

There is a universal consensus for the need for early detection of cardiac anomalies of the fetus amongst the obstetricians. Small complicated structure of a rapidly moving fetal heart posed insurmountable problems initially but persistence seems to pay off finally. The new hybrid technologies like the Bayesian Deep learning with Bayesian filter and deep learning intelligent neural network has been used for fetal ECG signal processing. Accuracies of over 90% is reported for ANN (Figure 1) and ELM for Cardiotocography [1,2]. Standard fetal ultrasound planes includes upper abdomen, 4 chamber view, 5 chamber view, short axis view, 3 vessel trachea and if needed longitudinal planes of aortic arch, the ductal arch and systemic veins.

ECCY-02-SI-02.pdf (ecronicon.com)