September 2, 2022Open Access

Abnormal ECG detection based on an adversarial autoencoder

Q: What is the clinical evidence from this study?

Study design: Other. Population: Abnormal ECG / Arrhythmia (n=44220). Intervention: ECG-AAE (Adversarial Autoencoder and Temporal Convolutional Network) vs. Other outlier detection methods. Primary outcome: Accuracy of abnormal ECG detection on the MIT-BIH dataset.

Key Result

The ECG-AAE framework detected abnormal ECGs with an accuracy of 0.9673 and AUC of 0.9672 on the MIT-BIH dataset, outperforming other popular outlier detection methods.

Structured PICO

Does the ECG-AAE framework improve the detection of abnormal ECGs compared to other outlier detection methods in ECG datasets?

Population

44,220 subjects from the MIT-BIH and CMUH datasets whose ECG data were used to train and test an adversarial autoencoder for abnormal ECG detection.

Intervention

ECG anomaly detection framework (ECG-AAE) based on an adversarial autoencoder and temporal convolutional network (TCN)

Comparator

13 other outlier detection methods including AnoGAN, AE, VAE, Stack LSTM, GRU, RNN, DEEP-SVDD, AE+OCSVM, DAGMM, GMM, OCSVM, iForest, and LOF

Outcome

Accuracy, precision, recall, F1-score, and AUC for detecting abnormal ECGssurrogate

The ECG-AAE framework, trained only on normal ECG data, effectively detects abnormal ECG signals and outperforms other popular outlier detection methods.

Limitations

False positive data are largely affected by noise interference
False negative data exist when there is baseline drift
The model may not perform ideally on ECG data with large variation such as patient movement

Abstract

Automatic detection and alarm of abnormal electrocardiogram (ECG) events play an important role in an ECG monitor system; however, popular classification models based on supervised learning fail to detect abnormal ECG effectively. Thus, we propose an ECG anomaly detection framework (ECG-AAE) based on an adversarial autoencoder and temporal convolutional network (TCN) which consists of three modules (autoencoder, discriminator, and outlier detector). The ECG-AAE framework is trained only with normal ECG data. Normal ECG signals could be mapped into latent feature space and then reconstructed as the original ECG signal back in our model, while abnormal ECG signals could not. Here, the TCN is employed to extract features of normal ECG data. Then, our model is evaluated on an MIT-BIH arrhythmia dataset and CMUH dataset, with an accuracy, precision, recall, F1-score, and AUC of 0.9673, 0.9854, 0.9486, 0.9666, and 0.9672 and of 0.9358, 0.9816, 0.8882, 0.9325, and 0.9358, respectively. The result indicates that the ECG-AAE can detect abnormal ECG efficiently, with its performance better than other popular outlier detection methods.

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper

اسأل الذكاء الاصطناعي

Bookmark

View Full Paper

Cite This Study

Shan et al. (Fri,) conducted a other in Abnormal ECG / Arrhythmia (n=44,220). ECG-AAE (Adversarial Autoencoder and Temporal Convolutional Network) vs. Other outlier detection methods was evaluated on Accuracy of abnormal ECG detection on the MIT-BIH dataset. The ECG-AAE framework detected abnormal ECGs with an accuracy of 0.9673 and AUC of 0.9672 on the MIT-BIH dataset, outperforming other popular outlier detection methods.

synapsesocial.com/papers/6a372a8cd9e0d8cad3c75b58 https://doi.org/https://doi.org/10.3389/fphys.2022.961724

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: