December 16, 2022Applied SciencesOpen Access

The average classification accuracy of the APRN on the four datasets was 97.89%, compared to 97.17% for CNN, 97.53% for CNN-R, and 97.73% for ResNet-18.

View Full Paper

Ask AI

Bookmark

Why the study?

Establishing an effective deep learning model to assist physicians in early screening and identifying ECG abnormalities can improve diagnostic accuracy.

Does the alternate pooling residual network (APRN) model improve ECG classification accuracy compared to standard CNN and ResNet models in standard ECG databases?

Population

ECG datasets from American MIT-BIH arrhythmia and ST segment abnormality, European ST-T, and sudden cardiac death databases

Comparison

Alternate pooling residual network vs CNN, CNN-R, and ResNet-18

Design

Algorithm development and comparative validation study

Authors

JZJunbin ZangNorth University of China JWJuliang WangHuazhong University of Science and Technology ZZZhidong ZhangHebei Agricultural University

Discussion

Member takes

Overview

APRN may marginally improve ECG classification accuracy; hypothesis-generating and requires prospective clinical validation before adoption.

Structured PICO

Does the alternate pooling residual network (APRN) model improve ECG classification accuracy compared to standard CNN and ResNet models in standard ECG databases?

Population

ECG datasets from the American MIT-BIH arrhythmia and ST segment abnormality database, European ST-T database, and sudden cardiac death ambulatory ECG database

Intervention

Alternate pooling residual network (APRN) model with wavelet adaptive threshold denoising algorithm

Comparator

Convolutional neural network (CNN), CNN with one residual unit (CNN-R), and deep residual network (ResNet-18)

Outcome

Classification accuracy of ECG signalssurrogate

The proposed alternate pooling residual network (APRN) model demonstrates high accuracy in classifying ECG abnormalities, outperforming standard CNN and ResNet-18 models.

Cite This Study

Zang et al. (2022) studied this question.

synapsesocial.com/papers/6a1a7c3e77ec05d9a7b89a83 https://doi.org/10.3390/app122412957

View Full Paper

Ask AI

Bookmark

December 16, 2022Applied SciencesOpen Access

The average classification accuracy of the APRN on the four datasets was 97.89%, compared to 97.17% for CNN, 97.53% for CNN-R, and 97.73% for ResNet-18.

View Full Paper

Ask AI

Bookmark

Why the study?

Establishing an effective deep learning model to assist physicians in early screening and identifying ECG abnormalities can improve diagnostic accuracy.

Does the alternate pooling residual network (APRN) model improve ECG classification accuracy compared to standard CNN and ResNet models in standard ECG databases?

Population

ECG datasets from American MIT-BIH arrhythmia and ST segment abnormality, European ST-T, and sudden cardiac death databases

Comparison

Alternate pooling residual network vs CNN, CNN-R, and ResNet-18

Design

Algorithm development and comparative validation study

Authors

JZJunbin ZangNorth University of China JWJuliang WangHuazhong University of Science and Technology ZZZhidong ZhangHebei Agricultural University

Discussion

Member takes

Overview

APRN may marginally improve ECG classification accuracy; hypothesis-generating and requires prospective clinical validation before adoption.

Structured PICO

Does the alternate pooling residual network (APRN) model improve ECG classification accuracy compared to standard CNN and ResNet models in standard ECG databases?

Population

ECG datasets from the American MIT-BIH arrhythmia and ST segment abnormality database, European ST-T database, and sudden cardiac death ambulatory ECG database

Intervention

Alternate pooling residual network (APRN) model with wavelet adaptive threshold denoising algorithm

Comparator

Convolutional neural network (CNN), CNN with one residual unit (CNN-R), and deep residual network (ResNet-18)

Outcome

Classification accuracy of ECG signalssurrogate

The proposed alternate pooling residual network (APRN) model demonstrates high accuracy in classifying ECG abnormalities, outperforming standard CNN and ResNet-18 models.

Cite This Study

Zang et al. (2022) studied this question.

synapsesocial.com/papers/6a1a7c3e77ec05d9a7b89a83 https://doi.org/10.3390/app122412957

View Full Paper

Ask AI

Bookmark