Automatic Detection of ECG Abnormalities by Using an Ensemble of Deep Residual Networks with Attention

Liu, Yang; Rui, Hebing; Wang, Kuanquan; Li, Qince; Sun, Qiang; Zhao, Na; Zhang, Henggui

doi:10.1007/978-3-030-33327-0_11

Cited by 5 publications

(5 citation statements)

References 9 publications

(19 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the Physionet Challenge 2017 competition (Clifford et al, 2017), Hannun's CNN based model (Hannun et al, 2019) achieved the best score. The DenseCNN wang (Wang et al, 2019a), and resnet1d Liu (Liu et al, 2019) achieved the top performance in China ECG AI Contest 2019 competition. (Zhao et al, 2020) proposed a large kernel size model SEresnet Zhao based on SE-block (Hu et al, 2018), achieving second place in the PhysioNet 2020 competition (Alday et al, 2020).…”

Section: Comparision With State-of-the-art Single Lead Ecg's Classifi...mentioning

confidence: 99%

MVKT-ECG: Efficient Single-lead ECG Classification on Multi-Label Arrhythmia by Multi-View Knowledge Transferring

Qin¹,

Li²,

Chen³

et al. 2023

Preprint

View full text Add to dashboard Cite

The widespread emergence of smart devices for ECG has sparked demand for intelligent singlelead ECG-based diagnostic systems. However, it is challenging to develop a single-lead-based ECG interpretation model for multiple diseases diagnosis due to the lack of some key disease information. In this work, we propose inter-lead Multi-View Knowledge Transferring of ECG (MVKT-ECG) to boost single-lead ECG's ability for multilabel disease diagnosis. This training strategy can transfer superior disease knowledge from multiple different views of ECG (e.g. 12-lead ECG) to single-lead-based ECG interpretation model to mine details in single-lead ECG signals that are easily overlooked by neural networks. MVKT-ECG allows this lead variety as a supervision signal within a teacher-student paradigm, where the teacher observes multi-lead ECG educates a student who observes only single-lead ECG. Since the mutual disease information between the single-lead ECG and muli-lead ECG plays a key role in knowledge transferring, we present a new disease-aware Contrastive Lead-information Transferring(CLT) to improve the mutual disease information between the single-lead ECG and muli-lead ECG. Moreover, We modify traditional Knowledge Distillation to multi-label disease Knowledge Distillation (MKD) to make it applicable for multi-label disease diagnosis. The comprehensive experiments verify that MVKT-ECG has an excellent performance in improving the diagnostic effect of single-lead ECG.* Equal contribution 1 AUTHORERR: Missing \icmlaffiliation.. AUTHORERR: Missing \icmlcorrespondingauthor.

show abstract

Section: Comparision With State-of-the-art Single Lead Ecg's Classifi...mentioning

confidence: 99%

MVKT-ECG: Efficient Single-lead ECG Classification on Multi-Label Arrhythmia by Multi-View Knowledge Transferring

Qin¹,

Li²,

Chen³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The convolutional operation in a CNN model can effectively extract morphological features from a raw noisy 1D or 2D data (Kiranyaz et al, 2019). In the improved version of CNN, residual convolutional neural network (ResNet) overcomes the degradation problem in DNNs by adding shortcut links between its layers (Liu et al, 2019). Therefore, we employed ResNet for human authentication using raw ECG waveform.…”

Section: The Proposed End-to-end Structure Resnet Modelmentioning

confidence: 99%

ResNet‐Attention model for human authentication using ECG signals

et al. 2020

Self Cite

View full text Add to dashboard Cite

Authentication is the process of verifying the claimed identity of the user. Recently, traditional authentication methods such as passwords, tokens, and so on are no longer used for authentication as they are more prone to theft and different types of violations. Therefore, new authentication approaches based on biometric modalities such as heartbeat pattern obtained from electrocardiogram (ECG) signals are considered. Unlike other biometrics, ECG provides the assurance that the person is alive, and is considered as one of the most accurate recent methods for authentication. In this article, two end-to-end deep neural network models for ECG-based authentication are proposed. In the first model, a convolutional neural network (CNN) is developed and in the second model, a residual convolutional neural network (ResNet) with attention mechanism called ResNet-Attention is designed for human authentication. We have used 2-s duration ECG signals obtained from two ECG databases (Physikalisch-Technische Bundesanstalt [PTB] and Check Your Bio-signals Here initiative [CYBHi]) for authentication. Our proposed ResNet-Attention algorithmachieved an accuracy of 98.85 and 99.27% using PTB and CYBHi, respectively. The results obtained by our developed model show that the performance is better than existing algorithms and can be used in real-time authentication systems after the validation with more diverse ECG data.

show abstract

“…The first one is to weigh different features output by the recurrent layers. The second one is to apply one attention module to each label [21]. The last one is to combine with the backbone CNNs, such as the squeeze-and-excitation layers combined with ResNet in Method 2 [13] and the self-attention combined with DenseNet in Method 5.…”

Section: Design Of Deep Neural Networkmentioning

confidence: 99%

A community effort to assess and improve computerized interpretation of 12-lead resting electrocardiogram

Ding

Wang

Yang

et al. 2021

Med Biol Eng Comput

Self Cite

View full text Add to dashboard Cite

Computerized interpretation of electrocardiogram plays an important role in daily cardiovascular healthcare. However, inaccurate interpretations lead to misdiagnoses and delay proper treatments. In this work, we built a high-quality Chinese 12-lead resting electrocardiogram dataset with 15,357 records, and called for a community effort to improve the performances of CIE through the China ECG AI Contest 2019. This dataset covers most types of ECG interpretations, including the normal type, 8 common abnormal types, and the other type which includes both uncommon abnormal and noise signals. Based on the Contest, we systematically assessed and analyzed a set of top-performing methods, most of which are deep neural networks, with both their commonalities and characteristics. This study establishes the benchmarks for computerized interpretation of 12-lead resting electrocardiogram and provides insights for the development of new methods.

show abstract

Automatic Detection of ECG Abnormalities by Using an Ensemble of Deep Residual Networks with Attention

Cited by 5 publications

References 9 publications

MVKT-ECG: Efficient Single-lead ECG Classification on Multi-Label Arrhythmia by Multi-View Knowledge Transferring

MVKT-ECG: Efficient Single-lead ECG Classification on Multi-Label Arrhythmia by Multi-View Knowledge Transferring

ResNet‐Attention model for human authentication using ECG signals

A community effort to assess and improve computerized interpretation of 12-lead resting electrocardiogram

Contact Info

Product

Resources

About