The main objective of F2ACCESS (RFID and Face-based Secure Event Access) is to open several lines of research to address the problem of automatic face recognition in highly complex scenarios, such as access to conferences and events, where exceptional security measures are required due to the diplomatic or institutional nature of the event, the profile of participants, or a high police security risk. The scenario in which this project is framed is characterized by uncontrolled lighting conditions, where users do not act collaboratively, and the captured data must be managed in real time. F2ACCESS is focused on acquiring the necessary technological knowledge in facial biometrics to increase the security level of access systems based on radio frequency identifiers (RFID), proposing a combined system of a physical card with biometric data, which are personal and non-transferable.
The proposed project belongs to the cooperation modality, with the following participants:
- Innova y Desarrolla en Red, S.L. (Desarrollanet)
- University of Vigo
Gradiant, subcontracted by Desarrollanet, is responsible for designing and implementing the robust facial recognition system in realistic environments.
To undertake this ambitious research project, various partial objectives must be covered. Regarding automatic face recognition, it is important to note that a realistic and highly non-collaborative scenario is intended to be addressed, where individuals walk toward an access door while biometric data is being captured. The following objectives can be identified: i) Robust detection and tracking of faces in motion, ii) Proper handling of detected faces to deal with changes in rotation angles that inevitably occur due to the relative position between the camera(s) and the person being recognized, iii) Use of technologies that allow robust recognition in varying lighting conditions, iv) Proper handling of appearance changes due to the presence/absence of a beard, mustache, or glasses, as well as changes in facial expression, and v) Real-time operation of the recognition system to allow smooth access. The special characteristics of the scenario considered in F2ACCESS distinguish this project significantly from previous work in facial recognition.
On the other hand, communication between the facial recognition system and the radio frequency-based system is crucial, as the fusion of both technologies will increase security even under the most unfavorable conditions. To achieve this, it will be necessary to synchronize the data from both subsystems, taking the necessary measures to verify that they come from the same individual. Additionally, data processing and management, both biometric and radio frequency, must be done in real time, imposing significant restrictions on handling the captured video stream, as well as efficient access to databases and quick decision-making. This way, security personnel will be automatically informed of possible alerts or unauthorized access.
