Enhancing physionet electrocardiogram records for fetal heart rate detection algorithm

The noninvasive fetal electrocardiogram (ECG) data available from Physionet data bank are suitable for developing fetal heart rate (FHR) detection algorithms. The data have been collected from single subject with a broad range of gestation weeks, and have a total data length of more than 9 hours arranged in 55 data sets. However, there are three additional data features which are currently not directly available from Physionet to facilitate the easy usage of these data: (1) the fetal peak visibility evaluation, (2) the gestation week, and (3) the data length. This article presents an improvement to the data bank by providing the additional features. The required pre-processing of the data is also discussed. This new data features are presented in bar graphs, and were used to assist the authors' two data channels-based FHR detection algorithm development. This detection scheme uses only one thoracic channel and one abdominal channel from this six Physionet data channels. The discussed data features and the data pre-processing would also be useful for researchers intending to use other channels combinations of the Physionet data.