Biometric recognition for a long time has been used in confined spaces, usually indoor, where security-critical operations required high accuracy recognition systems, e.g. in police stations, banks, companies, airports. Field activities, on the contrary, required more portability and flexibility leading to the development of devices for less constrained biometric traits acquisition and consequently of robust algorithms for biometric recognition in less constrained conditions. However, the application of "portable" biometric recognition, was still limited in specific fields e.g. for immigration control, and still required dedicated devices.
A further step would be to spread the use of biometric recognition on personal devices, as personal computers, tablets and smartphones. Some attempts in this direction were made embedding fingerprint scanners in laptops or smartphones. So far biometric recognition on personal devices has been employed just for a limited set of tasks, as to unlock the screen using fingerprints instead of passwords.
The research activities described in this thesis were focused on studying and developing solutions for iris recognition on mobile devices. This topic has been analyzed in all its main phases:
Acquisition: collection of the MICHE database, containing pictures of irises acquired by mobile devices;
Segmentation: development of an innovative iris segmentation algorithm;
Feature extraction and matching: iris recognition has been combined with the face and with sensor (smartphone) recognition.
Finally, the use of gaze analysis for human recognition has been investigated in order to verify its possible fusion with iris recognition.