Author : Ramesh A 1
Date of Publication :10th April 2023
Abstract: The commonality, measurement, individuality, and permanence of all biometric traits make them very desirable than any other forms of authentication. Numerous applications use biometric authentication, including unlocking phones using face recognition and fingerprint verification, using speech recognition to log into bank accounts, etc. . In a very similar way, the electrocardiogram(ECG) signals can potentially be employed as a biometric in order to identify a person. Because the ECG signals are specific to each and every individual and show how alive a person is. This undertaking offers a technique for user authentication using ECG signals. In this project, signals are processed and trained with Siamese Networks, and then models of trained Siamese Networks are used to identify persons. In this research, the Siamese Networks are trained using the ECG-ID dataset taken from the Physionet website in order to recognize specific individuals. 90 people’s raw ECG signals can be found in the ECG-ID dataset. The findings of this study can substantiate the claim that ECG signals can be used as biometrics.
Reference :