| Prediction of Heart Disease Using New Proposed CNN Model Architecture |
| Paper ID : 1005-ICEEM2023 (R4) |
| Authors |
|
Shaimaa Mahmoud *1, Mohamed Gaber2, Gamal Farouk2, Arabi Keshk2 1Computer Science Department, Faculty of Computers and Information, Menoufia University,Shebin Elkom 32511, Egypt 2Computer Science Department, Faculty of Computers and Information, Menoufia University,Shebin Elkom 32511, Egypt |
| Abstract |
| Classification of the electronic heart signals is one of the most important technologies used for the early detection of heart patients. In the context of reducing the average mortality rate, many contributions are made by researchers on the issue of the early detection of heart patients. However, the used methods were not accurate enough and required further improvement to get better performance. This paper introduces a new proposed model for the early prediction of heart disease patients based on the concept of deep convolutional neural networks (CNNs). This model is built of 6 convolutional layers, 3 MaxPooling layers, and 3 fully connected layers to improve the capacity of the prediction process. The used dataset contains 928 ECG images and was divided into four classes: normal, abnormal, history of myocardial infarction, and infarction patients. The proposed model selects the most important features from the ECGs images dataset. Compared with other CNN models such as LeNet-5, VGG-16, and VGG-19, the new proposed model achieved the best performance. The experimental results showed that the new proposed CNN model based on the ECG images dataset achieved an accuracy of 98%. |
| Keywords |
| Heart disease, Prediction, ECG, CNNs. |
| Status: Accepted |