HeartbeatID

The goal of the project is the development of biometric recognition algorithms based on the subject's heartbeat. The necessary electrocardiogram (ECG) signals will be acquired with an especially designed vest, and we will aim at achieving the lowest error rates under different physical activities, without compromising the privacy of the subject.

Project Description

In the ever-increasing complexity of the cybersecurity landscape, a reliable person authentication has become of the utmost importance. Biometric systems offer a number of advantages with respect to traditional authentication systems based on PINs or passwords, and have been deployed in different scenarios over the last decades.

Whereas attackers can easily produce artifacts to replicate one's face or fingerprints, and potentially use them to get illegal access to protected resources, replicating a person's heartbeat is not so straightforward. The main challenge remains, however, how to measure the heartbeat in a reliable but non-intrusive manner. Furthermore, external factors such as the environment (e.g., extreme low or high temperatures) or physical activities (e.g., having a relaxed conversation vs. doing sport) can have a big impact on the heart rate. Last but not least, biometric data are sensitive personal data which need to be protected - and, in the case of electrocardiogram (ECG) signals, further health-related information could be extracted from the ECG. This project will thus address those questions and focus on the development of robust, reliable, privacy-friendly biometric recognition systems based on ECG-signals captured with an specially designed vest.