Last week Ben Lagden and I attended Identity Week 2019 at the Excel in London. We were there to evaluate the technological advances the market was making in biometrics and more specifically facial recognition, as well as the associated issues with live presence detection or spoof detection.
The show was well attended and had the usual, established companies exhibiting, as well as some new and potentially disruptive technology businesses. Having spoken to a number of exhibitors, and now with the added benefit a few days of reflection allows, I feel there are two main schools of thought in terms of ensuring the capture of facial biometric data is live and actually captured at the expected time i.e. to prevent photograph or video replay attacks to fool the system. The first, uses a series of advanced algorithmic techniques to detect micro movement or dedicated hardware such as stereoscopic cameras or unique coloured light emitted from the LCD screen for the expected moment of capture. The second, simpler, technique requires asking the user to turn their head and look up/down. Both these solutions are potentially valid, but range hugely in both price and performance.
At Grosvenor Technology we are investing heavily in developing this technology, so we are able to deliver facial detection, verification and identification services to our customers, with the lowest possible FAR’s and FRR’s. In addition, we are also rapidly expanding our range of fingerprint-based biometric solutions, using world class solutions such as Lumidigm, Suprema and on clock identification of N:10,000+ utilising Innovatrics SDK’s. What’s really important is that the overall user experience remains seamless whilst preventing false acceptance, false rejection and spoof attacks whilst ensuring time we spend developing this technology doesn’t drive the cost of our time clocks out of reach for our partners.