Comparison of acceleration python library on design and implementation of QRS detection module from ECG heart signal

• Published in: IOP conference series. Materials Science and Engineering 2019-12, Vol.673 (1), p.12055
• Main Authors: Wicaksana, C A, Febriani, R, Muhammad, F
• Format: Article
• Language: English
• Subjects: Algorithms; Arrays; Convolution; Electrocardiography; Feature extraction; Hypertension; Libraries; Modules; Python
• ISSN: 1757-8981; 1757-899X
• DOI: 10.1088/1757-899X/673/1/012055

Electrocardiogram (ECG) signal is one of importance signal from our body that sourced from heart. There are many benefits that can be obtained from ecg signals, for example can determine whether sleepy/stress or not, and several diseasses like arrhytmia, hypertension, heart failure, etc. In this research, we created sub-module for extraction feature for ecg signal using GPU Acceleration, but in this research only emphasis on QRS detection and peak detection from ecg signal. The input will be ecg Signals and the output will be array of peak in every row. There have been various research trying to extract or detect QRS and peak based on Pan-Tomkins Algorithm, but this research will make use python and will compare the acceleration using some library. The flow comprise five main step, (1) load ecg signal, (2) filtered ecg, (3) derivative from filtered ecg, (4) squaring from derivative ecg, (5) convolution squaring ecg, and (6) peak detection using Fiducial Mark. The overall module has been succesfully implemented and compared in python. The result show that computation using numpy is still better and faster for small array data. The Output of peak of array can be used to the next module.