Cardiac Diseases detection by multimodal cardiac data classification using CNN-ANN cascaded algorithm

This paper presents early heart disease detection considering multimodal physiological parameters classification using a cascaded combination of Convolutional Neural Network and Artificial Neural Network algorithms. It has been established that ECG along with other physiological parameters such as age, blood pressure, blood sugar, etc, may provide more comprehensive and accurate diagnostic information about heart functioning In this work, CNN is used to classify the ECG into seven classes, and the classification accuracy is found to be $\mathbf{9 8. 6 9 \%}$ These ECG classes are mapped into three categories namely, normal, moderate, and severe The classified ECG data along with the thirteen other physiological parameters are fed into the ANN model for binary classification of heart disease. The experimental results of ANN output demonstrate that the patients with normal ECG may have heart disease. This multimodal analysis will further reduce the possibility of False Negative conditions during the primary screening of heart disease.