UFACE will apply emerging facial biometric methods to access control for physical access and healthcare (patient records). Recognition by the face has advantages over fingerprints, iris, etc. It is natural, culturally acceptable and easy to use unlike passwords or tokens, which are forgotten, lost or shared. An existing biometrics, which meets low security needs, will be enhanced to work on a wider range of facial variation (age, physiological state, etc.). A full prototype physical access control systems will be developed and assessed and two healthcare mock-ups will be developed for assessment by captive end users using an existing medical network. Full demonstrator based on the most promising mock-up will be built around a standards-based API. The partners are a research lab, a technology transfer company with expertise in the method, a smartcard supplier, a medical network company with experience in systems for the elderly and disabled, an organisation using the network and an end-user hosting the physical access control site.
To develop and demonstrate user friendly secure access control in financial services and healthcare. A facial biometric will be combined with a smart-card to create a personalised token. This will be integrated into the emerging biometric interface standards. Application demonstrators will show the advantages over current systems: more secure, less risk of unauthorised access and abuse, increased user acceptance, reduced cost and inconvenience.
The precise objectives are:
1) to develop new methods for the modelling of variation in facial appearance. These will be provided as a set of software routines,
2) to improve reliability and ease of use,
3) understand implementation and administration issues, by workshops and feedback from users and qualitative assessment,
4) to understand user issues especially for the elderly and disabled by qualitative assessment of mock-ups with these types of end-user, and
5) to integrate the biometric into standardised API for ease of re-use by participation in the standards process and promotion of enhancements to standards.
In phase 1, the technology transfer company VAL will support the 2 application developers in the creation of a set of mock-ups for assessment with end-users. For healthcare, access to patient records by physicians and patients will be explored. Use by the elderly and disabled will be particularly studied based on the experience of the medical network partner BioTrast. A phase 1 feedback workshop will be arranged between all the partners to share experiences and focus future development. An existing method based on statistical models which are complete (all faces are modelled), specific (only faces are modelled) and mathematically well founded will be adapted to create new methods for reliable facial identity modelling and verification. In particular: models of legitimate changes (age, weight, cosmetics, physiological state); techniques to improve robustness due to occlusions (glasses, beards); exploiting distinguishing marks (moles, scars); and models of rate of change of appearance. A make-up artist will create new images which will be used to create and test these models. This will also create a database of face images in varying physiological and cosmetic states for use as a resource to the research community. The emerging biometric APIs standardiation activities will be monitored and a compatible generic interface will be produced to facilitate application development.
In phase 2 full demonstrators based on the most promising mock-up will be built using the API and enchanced methods. This will be assessed by the captive end users. Assessment will be by estimation of the false acceptance and true reject rates. Dissemination will be at trade fairs and performance competitions, and the trade press in each country. Exploitation will be via the existing distribution channel of each partner.
1. feedback from end users on prototype and mock-ups developed using existing facial biometrics,
2. assessment from users of functional prototypes using enhanced methods.
1. new methods for reliable facial identity modelling and verification,
2. database of face images,
3. a facial recognition API compatible with emerging standards thus reducing integrator risk,
4. two application prototypes,
5. evaluation reports,
6. enhanced awareness of applications and benefits of facial biometrics.
Fields of science
- medical and health scienceshealth scienceshealth care serviceseHealth
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringcontrol systems
- natural sciencescomputer and information sciencescomputer securityaccess control
- engineering and technologymaterials engineering
- natural sciencesmathematicsapplied mathematicsstatistics and probability
Call for proposalData not available
Funding SchemeCSC - Cost-sharing contracts
See on map
See on map
See on map
See on map
M13 9PL Manchester
See on map