Automatic classification of electrocardiogram signals based on transfer learning and continuous wavelet transform

Sun, Xiaoming; Liu, Pengfei; He, Zhishuai; Yang, Han; Su, Bochao

doi:10.1016/j.ecoinf.2022.101628

Cited by 9 publications

(18 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The ECG classification model was supported by the GANAD-DFCNTA and CNN algorithms. The formulations were based on interpatient ECG classification to align with the settings in existing works [ 11 , 12 , 13 , 14 , 15 , 16 ] in an apple-to-apple comparison. The architecture of the CNN is summarized as follows: layer 1—conv1D with kernel = 50, unit = 128, ReLU = 3, and strides = 3; layer 2—batch normalization; layer 3—maximum pooling with size = 2 and stride = 3; layer 4—conv1-D with kernel = 8, unit = 32, ReLU = 1, and strides = 1; layer 5—batch normalization; layer 6—maximum pooling with size = 2 and stride = 2; layer 7—conv1-D with kernel = 5, unit = 512, ReLU = 1, and strides = 1; layer 8—conv1-D with kernel = 3, unit = 128, ReLU = 1, and strides = 1; layer 9—fully connected layer; and layer 10—output layer.…”

Section: Benchmark Datasets and Performance Evaluationmentioning

confidence: 99%

“…Various works considered a knowledge transfer from a source model with different domains compared with the target domain. Transfer learning was performed to transfer knowledge from a pretrained model using the ImageNet database to the MIT-BIH arrhythmia database [ 11 ]. Before model construction, the continuous wavelet transform was used to transform the 1D ECG signals to 2D ECG signals.…”

Section: Introductionmentioning

confidence: 99%

“…After the studies of the existing works in transfer learning for ECG classifications [ 11 , 12 , 13 , 14 , 15 , 16 ] and distant transfer learning applications [ 17 , 18 , 19 , 20 ], the major research limitations were illustrated as follows: There were inadequacies in the performance evaluation and analysis in some existing works that did not employ cross-validation [ 11 , 12 , 13 , 14 , 17 , 18 , 19 ] and did not conduct ablation studies on the components of the algorithms [ 11 , 12 , 14 , 19 , 20 ]. There was a lack of research on distant transfer learning for ECG classifications.…”

Section: Introductionmentioning

confidence: 99%

“…There were inadequacies in the performance evaluation and analysis in some existing works that did not employ cross-validation [ 11 , 12 , 13 , 14 , 17 , 18 , 19 ] and did not conduct ablation studies on the components of the algorithms [ 11 , 12 , 14 , 19 , 20 ].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Enhancing Electrocardiogram Classification with Multiple Datasets and Distant Transfer Learning

Chui¹,

Gupta

Zhao

et al. 2022

Bioengineering

View full text Add to dashboard Cite

Electrocardiogram classification is crucial for various applications such as the medical diagnosis of cardiovascular diseases, the level of heart damage, and stress. One of the typical challenges of electrocardiogram classification problems is the small size of the datasets, which may lead to limitation in the performance of the classification models, particularly for models based on deep-learning algorithms. Transfer learning has demonstrated effectiveness in transferring knowledge from a source model with a similar domain and can enhance the performance of the target model. Nevertheless, the consideration of datasets with similar domains restricts the selection of source domains. In this paper, electrocardiogram classification was enhanced by distant transfer learning where a generative-adversarial-network-based auxiliary domain with a domain-feature-classifier negative-transfer-avoidance (GANAD-DFCNTA) algorithm was proposed to bridge the knowledge transfer from distant sources to target domains. To evaluate the performance of the proposed algorithm, eight benchmark datasets were chosen, with four from electrocardiogram datasets and four from the following distant domains: ImageNet, COCO, WordNet, and Sentiment140. The results showed an average accuracy improvement of 3.67 to 4.89%. The proposed algorithm was also compared with existing works using traditional transfer learning, revealing an average accuracy improvement of 0.303–5.19%. Ablation studies confirmed the effectiveness of the components of GANAD-DFCNTA.

show abstract

Section: Benchmark Datasets and Performance Evaluationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Enhancing Electrocardiogram Classification with Multiple Datasets and Distant Transfer Learning

Chui¹,

Gupta

Zhao

et al. 2022

Bioengineering

View full text Add to dashboard Cite

show abstract

“…The categorization of heartbeats is required for the detection and classification of arrhythmias and the recognition of the classes of successive heartbeats to establish the heart rhythm category. [2][7] [9] Human-based categorization on a beat-by-beat basis is a time-consuming and complex operation. As a result, the automation of ECG analysis is critical for detecting heart problems that require immediate medical attention in clinical scenarios and saving the cardiologist time and effort.…”

Section: Introductionmentioning

confidence: 99%

ECG Arrhythmia Classification Using Convolutional Neural Network

Thapliyal¹,

Verma²,

Kumar³

2022

IJETAE

View full text Add to dashboard Cite

This study provides a thorough analysis of earlier DL techniques used to classify the ECG data. The large variability among individual patients and the high expense of labeling clinical ECG records are the main hurdles in automatically detecting arrhythmia by electrocardiogram (ECG). The classification of electrocardiogram (ECG) arrhythmias using a novel and more effective technique is presented in this research. A high-performance electrocardiogram (ECG)-based arrhythmic beats classification system is described in this research to develop a plan with an autonomous feature learning strategy and an effective optimization mechanism, based on the ECG heartbeat classification approach. We propose a method based on efficient 12-layer, the MIT-BIH Arrhythmia dataset's five micro-classes of heartbeat types and using the wavelet denoising technique. Compared to state-of-the-art approaches, the newly presented strategy enables considerable accuracy increase with quicker online retraining and less professional involvement.

show abstract

Transcending Boundaries: Assessing Transfer Learning’s Effectiveness in ECG-Based Heart Disease Prediction

Nag,

Mondal,

Hassan

et al. 2024

Soft Computing in Industry 5.0 for Sustainability

View full text Add to dashboard Cite

Automatic classification of electrocardiogram signals based on transfer learning and continuous wavelet transform

Cited by 9 publications

References 33 publications

Enhancing Electrocardiogram Classification with Multiple Datasets and Distant Transfer Learning

Enhancing Electrocardiogram Classification with Multiple Datasets and Distant Transfer Learning

ECG Arrhythmia Classification Using Convolutional Neural Network

Transcending Boundaries: Assessing Transfer Learning’s Effectiveness in ECG-Based Heart Disease Prediction

Contact Info

Product

Resources

About