ArXiv Preprint
Wearable devices are increasingly used, thanks to the wide set of
applications that can be deployed exploiting their ability to monitor physical
activity and health-related parameters. Their usage has been recently proposed
to perform biometric recognition, leveraging on the uniqueness of the recorded
traits to generate discriminative identifiers. Most of the studies conducted on
this topic have considered signals derived from cardiac activity, detecting it
mainly using electrical measurements thorugh electrocardiography, or optical
recordings employing photoplethysmography. In this paper we instead propose a
BIOmetric recognition approach using Wearable Inertial Sensors detecting Heart
activity (BIOWISH). In more detail, we investigate the feasibility of
exploiting mechanical measurements obtained through seismocardiography and
gyrocardiography to recognize a person. Several feature extractors and
classifiers, including deep learning techniques relying on transfer learning
and siamese training, are employed to derive distinctive characteristics from
the considered signals, and differentiate between legitimate and impostor
subjects. An multi-session database, comprising acquisitions taken from
subjects performing different activities, is employed to perform experimental
tests simulating a verification system. The obtained results testify that
identifiers derived from measurements of chest vibrations, collected by
wearable inertial sensors, could be employed to guarantee high recognition
performance, even when considering short-time recordings.
Emanuele Maiorana, Chiara Romano, Emiliano Schena, Carlo Massaroni
2022-10-18
