Deep Learning Approaches for Enhancing Cardiovascular Risk Assessment

Advanced diagnostic technologies are necessary for early diagnosis and treatment of cardiovascular disease (CVD), a rising global health concern. By combining raw ECG signals with patient demographic information, the suggested approach enhances the prediction of heart illness. To improve the system's anticipated accuracy, this method incorporates signal processing, feature extraction, and selection processes. Thus, we create an application that uses 12 lead-ECG signal images from PTB-XL dataset along with additional parameters like age, gender, height, and weight to forecast the vulnerability of a heart condition. This system method derives valuable properties from ECG data using a new Fast Fourier Transform (FFT) feature extraction algorithm. Furthermore, the random forest classification technique is used to select highly discriminative characteristics for prediction. Acquired features are input into a Multi-Layer Perceptron (MLP) model, which is trained by employing an extensive dataset of patient demographics and ECG signals.