Abstract:To a good extent, remote health monitoring (RHM) has become an important area of research as life expectancy has improved. RHM can be very helpful for chronic and critical disease patients. The security of patient identity as well as his or her health records (HR) is an important aspect of RHM. The patient can be identified using electrocardiogram (ECG) as a biometric identifier as the ECG pattern for each person is unique. An ECG-based patient identification system is proposed in this work where first ECG artefacts are suppressed and then passed through a combination of filters, i.e., digital differentiator and guided smoothening, which are helpful in important feature extraction and noise removal. The differentiated ECG signal is decomposed using the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise Prabhat Patel Long Short Term Memory (LSTM) based classifier. The results of the proposed method are presented in terms of accuracy and compared with the state of the art methods, and it has been found that the proposed method outperforms the state of the art methods.

SMHeart ID: smart ID based on ECG and long short term memory for remote health monitoring