Segment, Perceive and Classify - Multitask Learning of the Electrocardiogram in a Single Neural Network

Sodmann, Philipp; Vollmer, Marcus; Kaderali, Lars

doi:10.23919/cinc53138.2021.9662830

Cited by 3 publications

(3 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We applied 14 different methods to preprocess the ECGs: Corcodan [5] includes four steps: spike removal, low-pass Butterworth filter (60Hz cutoff), baseline drift removal using local linear estimates, substraction of trimmed average. From Neurokit2 [6] we applied: Neurokit (default) (high-pass 5th order Butterworth filter with 0.5 Hz cutoff, 50Hz powerline removal), pantomp-kins1985 (1st order Butterworth filters, bandpass cutoffs 5 Hz and 15 Hz [7]), hamilton2002 (1st order Butterworth filter, bandpass cutoffs 8 Hz and 16 Hz [8]), biosppy (FIR filter, cutoffs 3 Hz and 45 Hz [9]), elgendi2010 (2nd order Butterworth filter, bandpass cutoffs 8 Hz and 20 Hz [10]), engzeemod2012 (4th order Butterworth filter, bandstop cutoffs of 48 Hz and 52 Hz [11]).…”

Section: Preprocessing Methodsmentioning

confidence: 99%

Efficiency of different heartbeat detection methods by using alternative noise reduction algorithms

Vollmer¹,

Guzmán²

2022

Computing in Cardiology Conference (CinC)

View full text Add to dashboard Cite

Motivation: ECG noise reduction is an essential step in the ECG preprocessing pipeline. In particular, the quality of beat detection can be affected by several artifacts. The variety of available methods ranges from filter-based techniques (e.g. Butterworth or FIR), signal decomposition (e.g. wavelets) to neural networks (Cycle-GAN).Methods: 12-lead resting ECGs from the SFB/TR19 study on inflammatory dilated cardiomyopathy (n=704) and from the Study of Health in Pomerania (n=17,717) were preprocessed with 14 different methods to evaluate the accuracy of heartbeat detection methods in relation to the chosen method for preprocessing. Open source signal libraries (neurokit2, py-ecg-detectors, WFDB among others) with 34 detectors were evaluated. Sensitivity and Positive Predictive Values (PPV) were computed for each combination of preprocessing and detection method in a train/test scheme. Annotations were corrected for fixed delays and scored at a tolerance of 50 ms. Results and conclusion:Eplimited performed best, regardless of the chosen preprocessing method. For diseased ECGs, in Kalidas2017 it was seen an improvement in performance from 0.713 to 0.876 and from 0.778 to 0.879 for PPV and sensitivity respectively. Best results can be achieved with ECG leads V3, V5 and V6.

show abstract

Section: Preprocessing Methodsmentioning

confidence: 99%

Efficiency of different heartbeat detection methods by using alternative noise reduction algorithms

Vollmer¹,

Guzmán²

2022

Computing in Cardiology Conference (CinC)

View full text Add to dashboard Cite

show abstract

“…This study provides a classification to calculate specific properties, including PQ time, QTc, and Q-Q interval, within the network. In addition to the features derived during the calculations, the bottleneck layer in the U-Net network is proposed as an alternative for classification [12]. This study presents a fusion approach, associated with the DERMA dual event, as well as the fourier transform algorithm FrlFT in order to accurately identify normal and non-normal morphological properties in electrocardiogram signals [13].…”

Section: Related Literature 21 Deep Learning Tasks For Ecg Signal Cla...mentioning

confidence: 99%

Classification of electrocardiogram signals based on federated learning and a gaussian multivariate aggregation module

Meqdad

Hussein²,

Husain

et al. 2023

IJEECS

View full text Add to dashboard Cite

Categorization of cardiac abnormalities received from several centers is not possible within the quickest time because of privacy and security restrictions. Today, individuals’ security problem is considered as one of the most important research fields in most research sciences. This study provides a novel approach for detection of cardiac abnormalities based on federated learning (FL). This approach addresses the challenge of accessing data from remote centers and presents the possibility of learning without the need for transferring data from the main center. We present a novel aggregation approach in the FL for addressing the challenge of imbalanced data using the averaging stochastic weights (SWA) optimizer and a multivariate Gaussian in order to make a better and more accurate detection possible. The advantage of the present proposed approach is robust and secure aggregation for unbalanced electrocardiogram (ECG) data from heterogeneous clients. We were able to achieve 87.98% accuracy in testing with the robust VGG19 architecture.

show abstract

“…They also applied data augmentation by randomly cropping signals and randomly generating masks (Yang et al 2021). HeartlyAI applied different augmentation techniques such as cut-out, adding different types of noise, and allowing dropout of individual or groups of ECG channels (Sodmann et al 2021).…”

Section: Rank Team [Reference]mentioning

confidence: 99%

Issues in the automated classification of multilead ecgs using heterogeneous labels and populations

et al. 2022

View full text Add to dashboard Cite

Objective: The standard twelve-lead electrocardiogram (ECG) is a widely used tool for monitoring cardiac function and diagnosing cardiac disorders. The development of smaller, lower-cost, and easier-to-use ECG devices may improve access to cardiac care in lower-resource environments, but the diagnostic potential of these devices is unclear. This work explores these issues through a public competition: the 2021 PhysioNet Challenge. In addition, we explore the potential for performance boosting through a meta-learning approach. Approach: We sourced 131,149 twelve-lead ECG recordings from ten international sources. We posted 88,253 annotated recordings as public training data and withheld the remaining recordings as hidden validation and test data. We challenged teams to submit containerized, open-source algorithms for diagnosing cardiac abnormalities using various ECG lead combinations, including the code for training their algorithms. We designed and scored algorithms using an evaluation metric that captures the risks of different misdiagnoses for 30 conditions. After the Challenge, we implemented a semi-consensus voting model on all working algorithms. Main results: A total of 68 teams submitted 1,056 algorithms during the Challenge, providing a variety of automated approaches from both academia and industry. The performance differences across the different lead combinations were smaller than the performance differences across the different test databases, showing that generalizability posed a larger challenge to the algorithms than the choice of ECG leads. A voting model improved performance by 3.5%. Significance: The use of different ECG lead combinations allowed us to assess the diagnostic potential of reduced-lead ECG recordings, and the use of different data sources allowed us to assess the generalizability of algorithms to diverse institutions and populations. The submission of working, open-source code for both training and testing and the use of a novel evaluation metric improved the reproducibility, generalizability, and applicability of the research conducted during the Challenge.

show abstract

Segment, Perceive and Classify - Multitask Learning of the Electrocardiogram in a Single Neural Network

Cited by 3 publications

References 5 publications

Efficiency of different heartbeat detection methods by using alternative noise reduction algorithms

Efficiency of different heartbeat detection methods by using alternative noise reduction algorithms

Classification of electrocardiogram signals based on federated learning and a gaussian multivariate aggregation module

Issues in the automated classification of multilead ecgs using heterogeneous labels and populations

Contact Info

Product

Resources

About