Heterogeneous sensor data fusion acquisition model for medical applications

Abstract

Electrocardiogram (ECG) and portable technologies with Internet of Things (IoT)-based biometric authentication have recently gained popularity. As a cutting-edge, potent technique utilized in numerous ways to increase authentication effectiveness over the past few decades, ECG-based biometric verification has garnered a lot of attention. However, a user's ECG signal may alter based on their health or physical condition, which would prevent verification. It should be vital to create a trustworthy method that takes into account unique ECG variations for verification to be successful. An effective and trustworthy ECG check technique is provided in this study. Using the concept of domain customization, this study presents a novel supervised learning platform. Data from many systems has been combined into a unique feature and given to a particular grader, like a Support Vector Machine (SVM), for verification. Cross-validation searches on two accessible data sets were used to assess how effective the proposed verification scheme was. The evaluation results show that the effectiveness of our suggested system achieved a verification prediction performance of 99.4% with a high level of quality & memory. The conclusions showed that the suggested strategy was well suitable for actual-time software to execute the task.