PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
May 7, 2019Applied SciencesOpen Access

Myocardial Infarction Classification Based on Convolutional Neural Network and Recurrent Neural Network

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Wearable and portable ECG devices enable timely myocardial infarction detection, motivating the development of automated algorithms to classify myocardial infarction ECG signals without complex handcrafted features.

Population

ECG signals from the Physikalisch-Technische Bundesanstalt (PTB) database

Design

Algorithm development and validation study

Key result

A multi-channel automatic classification algorithm combining a 16-layer CNN and LSTM achieved an accuracy of 95.4%, sensitivity of 98.2%, and specificity of 86.5% for myocardial infarction ECGs.

Authors

KFKai FengXPXitian PiHLHongying Liu

Discussion

Loading...

Member takes

Overview

Does not yet support clinical adoption of CNN-LSTM MI classifiers; leaves open prospective validation.

Structured PICO

P
Population
ECG signals from the Physikalisch-Technische Bundesanstalt (PTB) database
I
Intervention
Multi-channel automatic classification algorithm combining a 16-layer convolutional neural network (CNN) and long-short term memory network (LSTM)
O
Outcome
Classification performance (accuracy, sensitivity, specificity, F1 score)

A combined CNN and LSTM algorithm can accurately classify myocardial infarction from I-lead ECG signals without requiring complex handcrafted features.

Cite This Study

Feng et al. (2019) studied Myocardial infarction. Multi-channel automatic classification algorithm combining a 16-layer CNN and LSTM was evaluated on Classification performance (accuracy, sensitivity, specificity, F1 score). A multi-channel automatic classification algorithm combining a 16-layer CNN and LSTM achieved an accuracy of 95.4%, sensitivity of 98.2%, and specificity of 86.5% for myocardial infarction ECGs.

synapsesocial.com/papers/6a70738cf44fa9f079de198fhttps://doi.org/10.3390/app9091879
View Full Paper
Ask AI
Bookmark
Share