Smart Human Identification System Based on PPG and ECG Signals in Wearable Devices

In this paper, we propose a novel privacy-preserving identification-permission system (entitled Smart Human Identification System) to provide human identification, privacy, and security for users of wearables devices. To perform this function, we divided the system into six algorithms. We used PPG and ECG signals from two public datasets (MIMIC and CapnoBase). The proposed system creates a human ID and then compares it to other individuals and tests its accuracy to evaluate the quality of using biosignals for direct human identification in wearable devices. The experimental results indicate that the proposed system presented accuracy for the PPG signals from MIMIC and CapnoBase equal to 96.875% and 99.15%, respectively. For the ECG signals from MIMIC and CapnoBase, the accuracy obtained was 96.6% and 90.66%, respectively.