فاطمه معاون

The electrocardiogram (ECG) is a recording of the changes in the bioelectrical potentials of the heart over the course of the cardiac cycle in humans. This approach provides crucial insights into the functional aspects of cardiac and cardiovascular systems. Timely identification of heart conditions and irregularities can lead to longer lifespan and enhanced life quality with appropriate medical intervention. As a result, numerous studies have been conducted on the analysis of ECG signals. To perform an effective diagnosis, it is necessary to examine ECG patterns and heartbeats over an extended period, which can take several hours. Consequently, the volume of data is significantly large, making the analysis computationally intensive and time-consuming. Additionally, there is a risk of misinterpretation or loss of vital information by the analyst. In this study, the analysis of electrocardiogram signals using neural networks has been explored. The primary goal of this study is to eva‎luate the performance of five neural networks (Convolutional Neural Network, Multilayer Perceptron, Long Short-Term Memory, SqueezeNet, and GoogLeNet) in classifying cardiac signals. The findings of this study demonstrated that the pre-trained GoogLeNet model achieved the best performance among these five neural networks, with an accuracy of 100% in classifying cardiac signals. This method also performed more successfully compared to previous methods.

الكتروكارديوگرام , طبقه‌بندي , تبديل موجك , شبكه‌هاي عصبي , اسكالوگرام , مقايسه

Electrocardiogram , classification , wavelet transform , neural networks , Scalogram , Comparison

Fatemeh Moaven

Dr.Mahbobeh molavi-arabshahi _ toraj nik-azad