May 21, 2019IEEE Journal of Biomedical and Health InformaticsOpen Access

LSTM-Based ECG Classification for Continuous Monitoring on Personal Wearable Devices

View Full Paper

Ask AI

Bookmark

Why the study?

Existing compute-intensive deep-learning approaches for ECG classification are unsuitable for continuous cardiac monitoring on wearable devices with limited processing capacity.

Does a novel LSTM-based ECG classification algorithm improve arrhythmia detection accuracy and execution time compared to previous algorithms on wearable devices?

Comparison

Wavelet transform and multiple LSTM algorithm vs previous works

Discussion

Member takes

Overview

Enables real-time arrhythmia detection on wearables; leaves open prospective clinical validation before practice adoption.

Structured PICO

Does a novel LSTM-based ECG classification algorithm improve arrhythmia detection accuracy and execution time compared to previous algorithms on wearable devices?

Population

ECG signals from the MIT-BIH ECG arrhythmia database (DS100 and DS200)

Intervention

Novel ECG classification algorithm employing wavelet transform and multiple long short-term memory (LSTM) recurrent neural networks

Comparator

Previous ECG classification algorithms (e.g., Hu et al., Chazal et al., Jiang and Kong, Ince et al., Kiranyaz et al.)

Outcome

ECG classification performance (Accuracy, Sensitivity, Specificity, Positive Predictivity, F1 score) and execution timesurrogate

A novel, lightweight LSTM-based ECG classification algorithm provides highly accurate, real-time arrhythmia detection suitable for continuous monitoring on personal wearable devices.

Cite This Study

A 2019 study studied this question.

synapsesocial.com/papers/6a7de94110b6f5f3737d96b2 https://doi.org/10.1109/jbhi.2019.2911367

View Full Paper

Ask AI

Bookmark

May 21, 2019IEEE Journal of Biomedical and Health InformaticsOpen Access

LSTM-Based ECG Classification for Continuous Monitoring on Personal Wearable Devices

View Full Paper

Ask AI

Bookmark

Why the study?

Existing compute-intensive deep-learning approaches for ECG classification are unsuitable for continuous cardiac monitoring on wearable devices with limited processing capacity.

Does a novel LSTM-based ECG classification algorithm improve arrhythmia detection accuracy and execution time compared to previous algorithms on wearable devices?

Comparison

Wavelet transform and multiple LSTM algorithm vs previous works

Discussion

Member takes

Overview

Enables real-time arrhythmia detection on wearables; leaves open prospective clinical validation before practice adoption.

Structured PICO

Does a novel LSTM-based ECG classification algorithm improve arrhythmia detection accuracy and execution time compared to previous algorithms on wearable devices?

Population

ECG signals from the MIT-BIH ECG arrhythmia database (DS100 and DS200)

Intervention

Novel ECG classification algorithm employing wavelet transform and multiple long short-term memory (LSTM) recurrent neural networks

Comparator

Previous ECG classification algorithms (e.g., Hu et al., Chazal et al., Jiang and Kong, Ince et al., Kiranyaz et al.)

Outcome

ECG classification performance (Accuracy, Sensitivity, Specificity, Positive Predictivity, F1 score) and execution timesurrogate

A novel, lightweight LSTM-based ECG classification algorithm provides highly accurate, real-time arrhythmia detection suitable for continuous monitoring on personal wearable devices.

Cite This Study

A 2019 study studied this question.

synapsesocial.com/papers/6a7de94110b6f5f3737d96b2 https://doi.org/10.1109/jbhi.2019.2911367

View Full Paper

Ask AI

Bookmark