1 Dimensional Residual Network for Ecg Signal Denoising

The electrocardiogram (ECG) is a medical examination that records the electrical activity of the heart. It helps healthcare professionals to detect various heart conditions, such as dysrhythmias. However, the existence of internal and external noise on the site renders precise diagnosis of these conditions difficult. Therefore, we opted for a 1 Dimensional Residual Network, which estimates the noise present in the corrupted ECG signal and subtracts it to produce a denoised output. The performance of the proposed model is then compared to that of Wavelet Transform (WT) and NonLocal Means (NLM) methods. The 1D ResNET demonstrated superior SNRout, highlighting its effectiveness in denoising ECG signals. Proving its potential to maintain signal integrity while effectively removing various sources of noise.