ECG De-noising Techniques and Optimal Feature Selection Using Principle Component Analysis

Abstract

Abstract

ECG (Electrocardiography) is used to record and determine the condition of the heart. This paper provides an overview of various ECG de-noising techniques that are used to eliminate different type of noises; therefore, noise reduction procedure to be performed to eliminate different type of noises such as baseline wander, dc offset and high frequency interference, and then the pre-processed signal is used to extract features from the ECG signal. This work shows performance comparison of various de-noising techniques. SNR (Signal to Noise Ratio) and RMSE (Root Mean Square Error) are the parameters used to measure the performance of the output signal. This paper also provides optimal and compressed features using principle component analysis (PCA). The present work was tested using a standard ECG dataset authorized by MIT-BIH. In this study, we have three stages: signal acquired, signal de-noising, and feature extraction using PCA.

Keywords: Electrocardiogram (ECG), ECG signal de-noising, wavelet de-noising, PCA, feature extraction

Cite this Article

Tamanna, Manoj Duhan. ECG De-noising Techniques and Optimal Feature Selection Using Principle Component Analysis. Journal of Telecommunication, Switching Systems and Networks. 2018; 5(1): 14–20p.