Loading [a11y]/accessibility-menu.js
Comparing two different types of deep neural networks to improve accuracy in ECG intepretation | IEEE Conference Publication | IEEE Xplore

Comparing two different types of deep neural networks to improve accuracy in ECG intepretation


Abstract:

Electrocardiogram(ECG) is an important method to investigate heart diseases. ECG remains one of the main investigations by current medical standards, playing an important...Show More

Abstract:

Electrocardiogram(ECG) is an important method to investigate heart diseases. ECG remains one of the main investigations by current medical standards, playing an important role in decision making for treating cardiac patients. Current ECG medical devices record 12 one-dimensional electrical signal channels to create a 3D map of the heart electrical activity. The modern ECG machines also offer physicians an automatic medical interpretation of the ECG based on measurement of the signal waves and waves intervals: PR, QRS, ST and T wave interval. We propose a novel computer assisted ECG interpretation method based on ID convolutional deep neural networks. This method is based on combining different types of neural networks layers: ID convolutional, Dense, Dropout, Flatten, MaxPooling. By using real electrocardiogram Data(having multiple types of real electrocardiogram data) from ECG recording we were able to improve the accuracy of the overall classification of our best deep neural network and this led to a better predictions of our deep neural network model. Our final results were promising, and we plan to further develop a system that can classify the electrocardiogram of a patient for current use in a hospital.
Date of Conference: 20-23 October 2021
Date Added to IEEE Xplore: 17 November 2021
ISBN Information:
Print on Demand(PoD) ISSN: 2372-1618
Conference Location: Iasi, Romania

References

References is not available for this document.