Implementation Paper on Contactless Heart-Beat Detection Using Image Processing

In recent years, cardiovascular health has become a major concern, with heart attacks being a leading cause of death. To address this, we propose an innovative application that predicts the likelihood of a heart attack within the next few hours. The application begins by having users log in and complete a brief assessment, collecting vital parameters such as age, diabetes, smoking, alcohol consumption, fitness level, and family medical history. Users then undergo a 60-second facial scan, where the Haar Cascade algorithm detects face and lighting changes. The scan captures subtle color shifts in blood vessels, which are processed using OpenCV and EVM technology to detect RGB color variations, providing critical data on blood flow. Using Fast Fourier Transform (FFT) analysis, the heart rate frequency is obtained, and the Heart Rate Variability (HRV) is calculated to determine stress levels and SpO2. This data, combined with the assessment, feeds into our machine learning model, which predicts the probability of a heart attack. The system generates graphical representations of key health parameters, and if the risk is critical, users can immediately contact a doctor through the app. With a 96% accuracy rate for heart attack prediction, we continue to refine the system to provide a reliable and accurate tool for heart health monitoring, empowering users to take timely medical action.