Fetal ECG extraction from abdominal recordings using array signal processing

An algorithm to automatically locate QRS complexes in noninvasive fetal ECG signals is described and was entered in the PhysioNet/CinC 2013 “Noninvasive Fetal ECG” challenge. The algorithm is based on an iterative subspace decomposition and filtering of the maternal ECG components from the recordings of a set of electrodes placed on the mother's abdomen. Once the maternal components are removed, a novel merging technique is applied to merge the recordings and generate a signal with a higher SNR to perform fetal peak detection. The algorithm produces an annotation file for each data set containing the location of the fetal QRS complexes in that set. The final results indicate that the algorithm is able to detect fetal peaks under different scenarios and for variety of devices and signals encountered in clinical practice.