Synthesis of standard 12‑lead electrocardiograms using two-dimensional generative adversarial networks

Zhang, Yuhe; Babaeizadeh, Saeed

doi:10.1016/j.jelectrocard.2021.08.019

Cited by 19 publications

(7 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…But, most previous ECG synthesis methods (Hossain et al, 2021;Hazra & Byun, 2020;Zhu et al, 2019;Kuznetsov et al, 2020;Thambawita et al, 2021;Kuznetsov et al, 2021) either yielded a single ECG view or independently synthesized multi-view ECG signals following various GAN models (Arjovsky et al, 2017) while omitting view correlations. Besides, most of the known methods (Zhu et al, 2019;Hazra & Byun, 2020) ignored the presence of heart diseases, while only a few (Zhang & Babaeizadeh, 2021;Golany et al, 2020) synthesized ECG signals belonging to different diseases using different models.…”

Section: Backgroundsmentioning

confidence: 99%

“…But, there were still concerns on developing automatic ECG classifiers, such as patient privacy protection (Hossain et al, 2021;Hazra & Byun, 2020) and absence of well-annotated ECG signals (Golany et al, 2020). Thus, a key upstream task (Golany et al, 2020;Zhang & Babaeizadeh, 2021) is to synthesize ECG signals in order to increase the diversity and quantity of training samples, which can also help reduce annotation needs if ECG signals are synthesized conditioned on heart diseases.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

ME-GAN: Learning Panoptic Electrocardio Representations for Multi-view ECG Synthesis Conditioned on Heart Diseases

Chen¹,

Liao²,

Ying³

et al. 2022

Preprint

View full text Add to dashboard Cite

Electrocardiogram (ECG) is a widely used noninvasive diagnostic tool for heart diseases. Many studies have devised ECG analysis models (e.g., classifiers) to assist diagnosis. As an upstream task, researches have built generative models to synthesize ECG data, which are beneficial to providing training samples, privacy protection, and annotation reduction. However, previous generative methods for ECG often neither synthesized multi-view data, nor dealt with heart disease conditions. In this paper, we propose a novel diseaseaware generative adversarial network for multiview ECG synthesis called ME-GAN, which attains panoptic electrocardio representations conditioned on heart diseases and projects the representations onto multiple standard views to yield ECG signals. Since ECG manifestations of heart diseases are often localized in specific waveforms, we propose a new mixup normalization to inject disease information precisely into suitable locations. In addition, we propose a view discriminator to revert disordered ECG views into a predetermined order, supervising the generator to obtain ECG representing correct view characteristics. Besides, a new metric, rFID, is presented to assess the quality of the synthesized ECG signals.Comprehensive experiments verify that our ME-GAN performs well on multi-view ECG signal synthesis with trusty morbid manifestations.

show abstract

Section: Backgroundsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

ME-GAN: Learning Panoptic Electrocardio Representations for Multi-view ECG Synthesis Conditioned on Heart Diseases

Chen¹,

Liao²,

Ying³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Zhang et al [13] proposed a GAN model, whose generator was comprised of a 2 dimensional BiLSTM plus a CNN layer. In the discriminator, they used CNN and FC layers.…”

Section: Related Workmentioning

confidence: 99%

“…Because of this, their use, even for scientific and research purposes, is highly regulated [8]. To address all these issues, the generation of synthetic ECG signals has been the focus of many studies ( [9], [10], [11], [12], [13], [14]).…”

Section: Introductionmentioning

confidence: 99%

Synthetic ECG Signal Generation Using Generative Neural Networks

Adib¹,

Afghah²,

Prevost³

2021

Preprint

View full text Add to dashboard Cite

Electrocardiogram (ECG) datasets tend to be highly imbalanced due to the scarcity of abnormal cases. Additionally, the use of real patients' ECGs is highly regulated due to privacy issues. Therefore, there is always a need for more ECG data, especially for the training of automatic diagnosis machine learning models, which perform better when trained on a balanced dataset. We studied the synthetic ECG generation capability of 5 different models from the generative adversarial network (GAN) family and compared their performances, the focus being only on Normal cardiac cycles. Dynamic Time Warping (DTW), Fréchet, and Euclidean distance functions were employed to quantitatively measure performance. Five different methods for evaluating generated beats were proposed and applied. We also proposed 3 new concepts (threshold, accepted beat and productivity rate) and employed them along with the aforementioned methods as a systematic way for comparison between models. The results show that all the tested models can to an extent successfully mass-generate acceptable heartbeats with high similarity in morphological features, and potentially all of them can be used to augment imbalanced datasets. However, visual inspections of generated beats favors BiLSTM-DC GAN and WGAN, as they produce statistically more acceptable beats. Also, with regards to productivity rate, the Classic GAN is superior with a 72% productivity rate.

show abstract

“…For related works, there have been attempts to classify A-fib using single-lead ECG signals and the data augmentation method 16 , 17 . Hatamian et al 18 proposed a method of classifying A-fib by combining the data augmentation and deep learning methods.…”

Section: Introductionmentioning

confidence: 99%

Classification feasibility test on multi-lead electrocardiography signals generated from single-lead electrocardiography signals

Yoon,

Joo

2024

Sci Rep

View full text Add to dashboard Cite

Nowadays, Electrocardiogram (ECG) signals can be measured using wearable devices, such as smart watches. Most wearable devices provide only a few details; however, they have the advantage of recording data in real time. In this study, 12-lead ECG signals were generated from lead I and their feasibility was tested to obtain more details. The 12-lead ECG signals were generated using a U-net-based generative adversarial network (GAN) that was trained on ECG data obtained from the Asan Medical Center. Subsequently, unseen PTB-XL PhysioNet data were used to produce real 12-lead ECG signals for classification. The generated and real 12-lead ECG signals were then compared using a ResNet classification model; and the normal, atrial fibrillation (A-fib), left bundle branch block (LBBB), right bundle branch block (RBBB), left ventricular hypertrophy (LVH), and right ventricular hypertrophy (RVH) were classified. The mean precision, recall, and f1-score for the real 12-lead ECG signals are 0.70, 0.72, and 0.70, and that for the generated 12-lead ECG signals are 0.82, 0.80, and 0.81, respectively. In our study, according to the result generated 12-lead ECG signals performed better than real 12-lead ECG.

show abstract

Synthesis of standard 12‑lead electrocardiograms using two-dimensional generative adversarial networks

Cited by 19 publications

References 22 publications

ME-GAN: Learning Panoptic Electrocardio Representations for Multi-view ECG Synthesis Conditioned on Heart Diseases

ME-GAN: Learning Panoptic Electrocardio Representations for Multi-view ECG Synthesis Conditioned on Heart Diseases

Synthetic ECG Signal Generation Using Generative Neural Networks

Classification feasibility test on multi-lead electrocardiography signals generated from single-lead electrocardiography signals

Contact Info

Product

Resources

About