Recurrence Plot-Based Approach for Cardiac Arrhythmia Classification Using Inception-ResNet-v2

Abstract

The present study addresses the cardiac arrhythmia (CA) classification problem using the deep learning (DL)-based method for electrocardiography (ECG) data analysis. Recently, various DL techniques have been utilized to classify arrhythmias, with one typical approach to developing a one-dimensional (1D) convolutional neural network (CNN) model to handle the ECG signals in the time domain. Although the CA classification in the time domain is very prevalent, current methods’ performances are still not robust or satisfactory. This study aims to develop a solution for CA classification in two dimensions by introducing the recurrence plot (RP) combined with an Inception-ResNet-v2 network. The proposed method for nine types of CA classification was tested on the 1st China Physiological Signal Challenge 2018 dataset. During implementation, the optimal leads (lead II and lead aVR) were selected, and then 1D ECG segments were transformed into 2D texture images by the RP approach. These RP-based images as input signals were passed into the Inception-ResNet-v2 for CA classification. In the CPSC, Georgia, and the PTB_XL ECG databases of the PhysioNet/Computing in Cardiology Challenge 2020, the RP-based method achieved an average F1-score of 0.8521, 0.8529, and 0.8862, respectively. The results suggested the excellent generalization ability of the proposed method. To further assess the performance of the proposed method, we compared the 2D RP-image-based solution with the published 1D ECG-based works on the same dataset. Also, it was compared with two traditional ECG transform into 2D image methods, including the time waveform of the ECG recordings and time-frequency images based on continuous wavelet transform (CWT). The proposed method achieved the highest average F1-score of 0.844, with only two leads of the 12-lead ECG original data, which outperformed other works. Therefore, the promising results indicate that the 2D RP-based method has a high clinical potential for CA classification using fewer lead ECG signals.

ECG Recurrence Plot-Based Arrhythmia Classification Using Two-Dimensional Deep Residual CNN Features

“…Table 16 compares the results of this study with those reported in other publications with respect to their respective arrhythmia categories, segment lengths, and average accuracies using various databases (mainly the MITDB). The Inception-ResNet-v2 network with RP images was used in Zhang et al [64] as a classification method for cardiac arrhythmias. The CPSC database detected nine types of arrhythmias in their proposed work.…”

Cardiac Arrhythmia classification based on 3D recurrence plot analysis and deep learning

“…Regarding the ECG-based experiments, the 12-lead ECG signals were transformed into 2DRP images and stacked together to form 3D images, as illustrated in Figure 4. The method of converting 1D ECG signals into the corresponding 2DRP images is reported in our previous work (Zhang et al, 2021). Then we applied with (min-max and z-score normalization) and without normalization to pre-process the 2D RPs, respectively, which are defined as follows.…”

“…After normalization, these 12 leads images were placed with the lead-index order of limb leads (lead I, II, III, aVR, aVL, aVF) followed by the chest leads (V1, V2, V3, V4, V5, V6) to form as a 3DRP image. In our previous work (Zhang et al, 2021), we used those 2D RP plots (see Figure 4) to train the network and detect 2D RP features for classification. The relationship between the leads is implicitly investigated by the network, which is essential to explore but less obvious to learn from the 2D textures.…”

“…To study nonlinear dynamic spatial features of the cardiac system for arrhythmia classification, the recurrence plot (RP) technique has been used to discover the recurrence pattern buried in the time series of ECG signals and then successfully applied to the detection of ventricular fibrillation, as well as the prediction of premature atrial complex, premature ventricular complex, and AF (Mathunjwa et al, 2021). In our recent work (Zhang et al, 2021), we successfully utilized the 2D RPs to distinguish various arrhythmias, leading to better solutions than linear approaches.…”

“…This work aims to develop further the RP technique into a 3D framework for improved arrhythmia classification. In our recent study (Zhang et al, 2021), the 2DRP images offer a unique feature detection mechanism for arrhythmia classification compared with conventional approaches. However, those 2DRP maps are directly fed into the neural network in a decoupled manner, without sorting and directly analyzing shared features and nonlinear alterations between these 2D images in the training process.…”