Surrey wins £6 million funding to develop superior facial recognition technology
Researchers from the University of Surrey’s Centre for Vision, Speech and Signal Processing (CVSSP) have secured a £6 million biometrics research programme grant from the Engineering and Physical Sciences Research Council.
The £6 million FACER2VM project will see researchers from the Centre for Vision, Speech and Signal Processing (CVSSP) within the Department of Electrical and Electronic Engineering lead the development of ‘unconstrained’ facial recognition technology, working alongside colleagues from Imperial College London and the University of Stirling.
The partnership aims to unlock the potential of recent advances in machine learning and 3D face and face dynamics modelling to extend automatic face recognition to unconstrained scenarios. Working with the Home Office’s Centre for Applied Science and Technology (CAST), they aim to develop technology capable of identifying faces captured in poor light, crowded areas, in images of low resolution and poor image quality.
Their five years of research is due to commence on 1 January 2016 and marks a major UK research collaboration between internationally leading experts in face biometrics and face video analysis, the Home Office, BBC and industry. The grant was awarded under the flagship funding mechanism of the EPSRC (for research teams that are recognised as world-leading), with CVSSP holding two such EPSRC Programme Grants.
Currently, facial recognition technology is limited in its applicability. UK airports began screening airline passengers with facial recognition technology as an alternative to the use of passport officers in 2008. This attempt to ease congestion and improve security demonstrated the scope of such technology but came with serious limitations, as subjects would have to stand still and look directly into the camera for the scans to be successful.
Surrey’s own Professor Josef Kittler, an expert in biometrics from CVSSP, will act as the principle investigator as FACER2VM works towards developing facial recognition technology free of these confines.
The key drivers behind the need to more effectively recognise or prove identity of individuals are the rapidly growing mobility of people and the changes in conducting business in the Digital Economy. It is also driven by continuous efforts to improve the security of the population by enhancing protection against fraud, crime and terrorism.
Surrey will work alongside Imperial College London regarding the biometrics, whilst the University of Stirling focus their efforts on the psychology of face perception to inspire and inform the engineering research.
Professor Josef Kittler commented “Speaking on behalf of Professor Maja Pantic, Imperial, Professor Peter Hancock, Stirling and myself, we are confident that this exciting collaborative multidisciplinary project will break new grounds in unconstrained face recognition for the benefit of security, forensics, digital economy and smart cities applications.”