Identification of Fetal QRS with Base of Abdominal Electrocardiogram using Backpropagation Artifical Neural Network

摘要

A method for fetal QRS identification (FECG) originating from a non-invasive recording of abdominal electrocardiogram (AECG) is conducted. The method utilizes a backpropagation Artificial Neural Network (ANN). An abdominal electrocardiogram recording is extracted to obtain an electrocardiographic QRS. The method is tested using the clinical dataset of Abdominal and Direct Fetal Electrocardiogram Database (adfecgdb) from Physionet. The database consists of patients in pregnancy. Four leads of an electrocardiogram are utilized. This study is conducted by varying features of the abdominal electrocardiogram, and the detection method shows a good performance.