ECGlib: Library for processing electrocardiograms

To facilitate evaluation of ECGs in the FDA ECG Warehouse, public databases and clinical trials we have developed a C++ library for processing ECGs (ECGlib). ECGlib has a modular design, and is capable of handling files stored in many different formats, e.g. ISHNE, Physionet and FDA HL7. Moreover, ECGlib provides functions to do standard ECG signal processing, such as noise removal, QRS detection, classification, median beat creation and ECG waveform delineation. The performance of the different components of ECGlib has been evaluated using publicly available databases from Physionet (MIT-BIH and QTDB). The performance of ECGlib processing methodologies is comparable to state-of-the-art methods. We have also developed a MATLAB/Octave interface for ECGlib and are working on an interface for Python, R and Julia. Lastly, ECGlib comes with a set of command line tools that utilize parallel processing to quickly enable researchers to process large databases. We believe frameworks such as the one described can be used to facilitate research of ECG signals and we are working on making the library publicly available under an open source license.