Classification of ECG signal using FFT based improved Alexnet classifier

Abstract

Electrocardiograms (ECG) are extensively used for the diagnosis of cardiac arrhythmias. This paper investigates the use of machine learning classification algorithms for ECG analysis and arrhythmia detection. This is a crucial component of a conventional electronic health system, and it frequently necessitates ECG signal reduction for long-term data storage and remote transmission. Signal processing methods must be used to extract the function of the morphological properties of the ECG signal changing with time, which is difficult to discern in the typical visual depiction of the ECG signal. In biomedical research, signal processing and data analysis are commonly employed methodologies. This work proposes the use of an ECG arrhythmia classification method based on Fast Fourier Transform (FFT) for feature extraction and an improved AlexNet classifier to distinguish the difference between four types of arrhythmia conditions that were collected from records. The Convolutional Neural Network (CNN) algorithm’s results are compared to those of other algorithms, and the simulation results prove that the proposed technique is more effective for various parameters. The final results of the proposed system show that its ability to find deviations is 20% better than that of traditional systems.

A review on Machine, Transfer and Deep learning approaches for ECG classification

“…High Order Statistics (HOS): exhibit good performance to describe ECG morphology [18]. each beat is splitted into five segments, creating a 10-dimensional feature, calculating the value of skewness and kurtosis value over each segment as illustrated in equations (2,3).…”

“…The Multiple layer Perceptron (MLP) [19] architecture is presented in figure 3, it consists of main three modules: the input layer, the hidden layer/s, and the output layer. MLP processes information due to the following equation, for each node: (16) Here, wij (1) and wkj (2) are the weights of the output and the hidden neurons, respectively, yi represents the output of the network, and f is the used activation function [20]. The training process of the model can be processed in forward-propagation or back-propagation to update the weights of the input features to the network.…”

“…The 12-lead signal gives a complete view of heart structure and function that helps detect abnormalities and diagnose different diseases efficiently and accurately. It also gives a precise description for each part of the heart about its function performance which is required to do the correct recognition and classification [2]. The mentioned problem had exposed the need for automatic classification of ECG signals to facilitate and accelerate the diagnosis process correctly and efficiently.…”

“…The classification process of ECG signals can be generally splitted into two phases, the first phase is data preparation and the second is the classification process itself [2].…”

“…There are various techniques to solve ECG classification problems in both data preparation and classification phases. In the data preparation phase, algorithms can be used for denoising signals for low-and high-frequency noise that is represented in the signal such as generative adversarial networks or different function domain transformation techniques where noise components are demolished or can be ignored such as different families of Fourier and Wavelet transforms [2,3]. Also, there are algorithms to clean the data and reduce their dimensions to expose the principal components for the classification stage, some of these algorithms are principal component analysis (PCA), independent component analysis (ICA), and linear discriminant analysis (LDA).…”