Arrhythmia is a condition where the heart beats irregularly, too fast, or too slow. It can
be caused by various factors such as heart disease, high blood pressure, or electrolyte imbalances.
Arrhythmia is defined as an irregularity in the heartbeat rhythm. Recently, many people were affected
by cardiac diseases, due to covid19 [1]. So, we proposed the deep learning model to automate the
detection and classification process to help the doctors for quick results. We use ECG signals that were
downloaded from the Physio-Net database as our input. The Dataset undergoes a process to change the
signal into a space with many dimensions using some basic functions. Using wavelet scattering, we
extract the characteristics. Wavelet scattering is a signal processing technique that decomposes a signal
into its constituent waveforms, allowing for the identification of patterns that may indicate arrhythmias.
Then converting 1D ECG Signal to 2D scalogram images by using Continuous Wavelet Transform
(CWT). Arrhythmia detection is done by using deep learning (Efficientnet-B0) model and compared it
with other models such as Efficientnet-B0 with 97.9%, Alexnet model gives 86.67%, Resnet50 by
97.78%, and Densenet with 89.26% accuracy.
