ECG identification and monitoring system

Abstract

Technologies emerging the world have raised new threats. They made the traditional identification modalities vulnerable and easy to counterfeit. Hence, more sophisticated biometrics become indis- pensable to protect the integrity of the devices and the sensitive data. In the last few years there has been a growing interest in electrocardiogram (ECG) due to its uniqueness. In this work, we developed an ECG identification system. We employed this system into a real world application while solving a second problem which is the digitization of the health system. More precisely, it is an ECG remote monitoring system with ECG based identification. Two datasets were used: the MIT-BIH Arrhythmia and our own collected dataset using our implemented hardware system. The collected data was denoised using a wavelet transform based filter. After segmenting the data and extracting the necessary fiducial features, we stored the generated templates. A template matching algorithm based on the correlation factor was considered to make the matching decisions. Finally, the whole process is presented to the user through a user-friendly android application. Preliminary experimental results indicate that the system achieved 87.5% to 90.90% accuracy.