Cardiac Abnormality Detection Based on an Ensemble Voting of Single-Lead Classifier Predictions

Aublin, Pierre; Ammar, Mouin Ben; Achache, Nina; Benahmed, Melina; Hichami, Aymane El; Barret, Michel; Fix, Jérémy; Oster, Julien

doi:10.23919/cinc53138.2021.9662824

Cited by 6 publications

(10 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Finally, we removed the signal segments beyond the first 15 s if they were available. Hence, we used short ECG segments and avoided big differences in the lasting time of the dataset samples, as previously done in Xiaoyu et al (2021), Wickramasinghe and Athif (2021), Aublin et al (2021).…”

Section: Signal Preprocessingmentioning

confidence: 99%

From 12 to 1 ECG lead: multiple cardiac condition detection mixing a hybrid machine learning approach with a one-versus-rest classification strategy

et al. 2022

View full text Add to dashboard Cite

Objective. Detecting different cardiac diseases using a single or reduced number of leads is still challenging. This work aims to provide and validate an automated method able to classify ECG recordings. Performance using complete 12-lead systems, reduced lead sets, and single-lead ECGs is evaluated and compared. Approach. Seven different databases with 12-lead ECGs were provided during the PhysioNet/Computing in Cardiology Challenge 2021, where 88 253 annotated samples associated with none, one, or several cardiac conditions among 26 different classes were released for training, whereas 42 896 hidden samples were used for testing. After signal preprocessing, 81 features per ECG-lead were extracted, mainly based on heart rate variability, QRST patterns and spectral domain. Next, a One-versus-Rest classification approach made of independent binary classifiers for each cardiac condition was trained. This strategy allowed each ECG to be classified as belonging to none, one or several classes. For each class, a classification model among two binary supervised classifiers and one hybrid unsupervised-supervised classification system was selected. Finally, we performed a 3-fold cross-validation to assess the system’s performance. Main results. Our classifiers received scores of 0.39, 0.38, 0.39, 0.38, and 0.37 for the 12, 6, 4, 3 and 2-lead versions of the hidden test set with the Challenge evaluation metric ( C M ). Also, we obtained a mean G -score of 0.80, 0.78, 0.79, 0.79, 0.77 and 0.74 for the 12, 6, 4, 3, 2 and 1-lead subsets with the public training set during our 3-fold cross-validation. Significance. We proposed and tested a machine learning approach focused on flexibility for identifying multiple cardiac conditions using one or more ECG leads. Our minimal-lead approach may be beneficial for novel portable or wearable ECG devices used as screening tools, as it can also detect multiple and concurrent cardiac conditions.

show abstract

Section: Signal Preprocessingmentioning

confidence: 99%

From 12 to 1 ECG lead: multiple cardiac condition detection mixing a hybrid machine learning approach with a one-versus-rest classification strategy

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The proposed model was inspired from the solution presented in Aublin et al (2021) during the official phase of the 2021 PhysioNet/Computing in Cardiology Challenge, which will be referred as the baseline (BL) later in the manuscript. The DL model was inputted single-lead ECGs for classification purpose.…”

Section: Architecturementioning

confidence: 99%

“…A single lead-agnostic DL model was trained, which is therefore independent from the lead position. The architecture with squeeze and excitation (SE) blocks was shown to have similar performance to a simple CNN architecture during cross-fold validation in Aublin et al (2021), which were not worth the additional parameters and training time required for its optimisation. These blocks were therefore removed, and the proposed architecture is described in table 2.…”

Section: Architecturementioning

confidence: 99%

“…The same loss as in Aublin et al (2021) was used in this study, and contained two terms: (i) The first one was a weighted binary cross-entropy. As for that previous submission, a mask was used to cancel the contribution of a class when it was absent from the database where the sample came from.…”

Section: Loss and Optimizermentioning

confidence: 99%

“…The 2021 PhysioNet/Computing in Cardiology Challenge was an extension from 2020, and aimed at assessing whether a subset of ECG leads were sufficient for an overall ECG classification task (Reyna et al 2021). Among the best entries for the 2021 challenge, ensembling of model predictions (Nejedly et al 2021) During the official phase of the 2021 PhysioNet/Computing in Cardiology Challenge, we proposed a singlelead based fully DL model with the final decision on the different lead subsets being based on an average voting of all available single-lead predictions (Aublin et al 2021). In this extension of our challenge work, we hypothesise that combining handcrafted features with our deep learning approach can improve model performance.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Predict alone, decide together: cardiac abnormality detection based on single lead classifier voting

et al. 2022

Self Cite

View full text Add to dashboard Cite

Objective: A classifier based on weighted voting of multiple single-lead based models combining deep learning representation and hand-crafted features was developed to classify 26 cardiac abnormalities from different lead subsets of short-term electrocardiograms (ECG). Approach: A two-stage method was proposed for the multilead prediction. First a lead-agnostic hybrid classifier was trained to predict the pathologies from single-lead ECG signals. The classifier combined fully automated deep learning (DL) features extracted through a convolutional neural network with hand-crafted features through a fully connected layer. Second, a voting of the single-lead based predictions was performed. For the 12-lead subset, voting consisted in an optimised weighting of the output probabilities of all available single lead predictions. For other lead subsets, voting simply consisted in the average of the lead predictions. Main Results: This approach achieved a challenge test score of 0.48, 0.47, 0.46, 0.46, 0.45 on the 12, 6, 4, 3, 2-lead subsets respectively on the 2021 Physionet/Computing in Cardiology challenge hidden test set. The use of an hybrid approach and more advanced voting layer improved some individual class classification but did not offer better generalization than our baseline fully DL approach. Significance: The proposed approach showed potential at correctly classifying main cardiac abnormalities and dealt well with reduced lead subsets.

show abstract

Abnormalities analysis of electrocardiogram signals by using artificial intelligence

Dhara,

Bhanja,

Khampariya

2024

Multimed Tools Appl

View full text Add to dashboard Cite

Cardiac Abnormality Detection Based on an Ensemble Voting of Single-Lead Classifier Predictions

Cited by 6 publications

References 10 publications

From 12 to 1 ECG lead: multiple cardiac condition detection mixing a hybrid machine learning approach with a one-versus-rest classification strategy

From 12 to 1 ECG lead: multiple cardiac condition detection mixing a hybrid machine learning approach with a one-versus-rest classification strategy

Predict alone, decide together: cardiac abnormality detection based on single lead classifier voting

Abnormalities analysis of electrocardiogram signals by using artificial intelligence

Contact Info

Product

Resources

About