A Two-Phase Multilabel ECG Classification Using One-Dimensional Convolutional Neural Network and Modified Labels

Bruoth, Erik; Bugata, Peter; Gajdos, David; Horvat, S.; Hudak, David E.; Kmečová, Vladimíra; Staňa, Richard; Stanková, M; Szabari, Alexander; Vozáriková, Gabriela

doi:10.23919/cinc53138.2021.9662878

Cited by 10 publications

(11 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section, we present the extended results and comparison of our proposed models. As the added value to Antoni et al (2021), we compare here the training datasets in terms of individual labels and we identify the labels which need to be predicted using more than two leads. We also present the scores achieved in PhysioNet/CinC Challenge 2021 and the prototype application available in the cloud.…”

Section: Resultsmentioning

confidence: 99%

“…Then, we created a base model using a deep 1D convolutional neural network, which we subsequently optimized for Challenge metrics and adapted to the different characteristics of the competition datasets. Compared to Antoni et al (2021), in this section we describe in more details a modified mapping of original labels into a single scheme and post-Challenge updates to our solution.…”

Section: Methodsmentioning

confidence: 99%

“…The 1D variant of ResNet50 (He et al 2016) is used as a backbone with a number of virtual channels decreased by a factor of 4, which means that the input ECG slice is mapped to the 512-dimensional space of latent factors. The overall architecture is shown in figure 1 (for more details see our paper (Antoni et al 2021)).…”

Section: Base Modelmentioning

confidence: 99%

“…One of the goals of PhysioNet/CinC Challenge 2021 was to determine whether reduced lead configurations would provide sufficient classification accuracy compared to standard 12-lead ECGs. According to the results given in Antoni et al (2021), the answer depends on the specific label. Figure 3 shows the performance of 12-lead, 4-lead, and 6-lead models for each label expressed by the F2-score.…”

Section: Will Two Do?mentioning

confidence: 99%

See 3 more Smart Citations

Automatic ECG classification and label quality in training data

et al. 2022

Self Cite

View full text Add to dashboard Cite

Objective: Within the PhysioNet/Computing in Cardiology Challenge 2021, we focused on the design of a machine learning algorithm to identify cardiac abnormalities from electrocardiogram recordings (ECGs) with a various number of leads and to assess the diagnostic potential of reduced-lead ECGs compared to standard 12-lead ECGs. Approach: In our solution, we developed a model based on a deep convolutional neural network, which is a 1D variant of the popular ResNet50 network. This base model was pre-trained on a large training set with our proposed mapping of original labels to SNOMED codes, using three-valued labels. In the next phase, the model was fine-tuned for the Challenge metric and conditions. Main results: In the Challenge, our proposed approach (team CeZIS) achieved a Challenge test score of 0.52 for all lead configurations, placing us 5th out of 39 in the official ranking. Our improved post-Challenge solution was evaluated as the best for all ranked configurations, i.e., for 12-lead, 3-lead, and 2-lead versions of the full test set with the Challenge test score of 0.62, 0.61, and 0.59, respectively. Significance: In addition to building the model for identifying cardiac anomalies, we provide a more detailed description of the issues associated with label mapping and propose its modification in order to obtain a better starting point for training more powerful classification models. We compare the performance of models for different numbers of leads and identify labels for which two leads are sufficient. Moreover, we evaluate the label quality in individual parts of the Challenge training set.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Base Modelmentioning

confidence: 99%

Section: Will Two Do?mentioning

confidence: 99%

See 2 more Smart Citations

Automatic ECG classification and label quality in training data

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…As the field of AI has developed, more advanced deep learning-based techniques, such as convolutional neural networks (CNNs) [ 3 ], long short-term memory (LSTM) networks [ 18 ], and transformer neural networks [ 19 , 20 ] have been explored for the analysis of ECG signals. The classification of ECG data has many important applications, including but not limited to diagnosis in the absence of a cardiologist, automatic verification of ECG reports, and helping medics in teaching [ 21 ].…”

Section: Introductionmentioning

confidence: 99%

Cardiac Arrhythmia Classification Using Advanced Deep Learning Techniques on Digitized ECG Datasets

Sattar,

Mumtaz,

Qadir

et al. 2024

Sensors

View full text Add to dashboard Cite

ECG classification or heartbeat classification is an extremely valuable tool in cardiology. Deep learning-based techniques for the analysis of ECG signals assist human experts in the timely diagnosis of cardiac diseases and help save precious lives. This research aims at digitizing a dataset of images of ECG records into time series signals and then applying deep learning (DL) techniques on the digitized dataset. State-of-the-art DL techniques are proposed for the classification of the ECG signals into different cardiac classes. Multiple DL models, including a convolutional neural network (CNN), a long short-term memory (LSTM) network, and a self-supervised learning (SSL)-based model using autoencoders are explored and compared in this study. The models are trained on the dataset generated from ECG plots of patients from various healthcare institutes in Pakistan. First, the ECG images are digitized, segmenting the lead II heartbeats, and then the digitized signals are passed to the proposed deep learning models for classification. Among the different DL models used in this study, the proposed CNN model achieves the highest accuracy of ∼92%. The proposed model is highly accurate and provides fast inference for real-time and direct monitoring of ECG signals that are captured from the electrodes (sensors) placed on different parts of the body. Using the digitized form of ECG signals instead of images for the classification of cardiac arrhythmia allows cardiologists to utilize DL models directly on ECG signals from an ECG machine for the real-time and accurate monitoring of ECGs.

show abstract