Smartphone Face Recognition Improved by Copying the Brain

Face recognition security on smartphones can be significantly improved if users store an ‘average’ photo of themselves, according to new research by scientists at the University of York.

A research team led by Dr David Robertson, of the Department of Psychology’s FaceVar laboratory at York, found that combining different pictures of the user, rather than a single ‘target’ image, leads to much better recognition across all kinds of daily settings. The research is published in the journal PLOS ONE.

The researchers examined the performance of the ‘face unlock’ system on Samsung Galaxy phones. They found that the system was generally very good at rejecting imposters, but that it often failed to recognise the genuine owner too. However, performance could be greatly improved, often to perfect levels, if users stored an ‘average’ of their own photos – formed by morphing together several different photos of the user.

The idea for this technique comes from studies of human face recognition. It is understood that the brain forms abstract representations of the faces it knows, and that people are very good at recognising their family and friends over a range of conditions. If it is possible to copy these representations, then automatic face recognition can improve. Using ‘average’ images is a simple way to do this.

This shows two faces used in the study.
Image 1: Normal use of the smartphone ‘face-unlock’ system, it stores a single image of the owner.
Image 2: Shows the owner’s face ‘average’ – an image derived from many photos of a person – being encoded into the system. Image adapted from the University of York press release.

Dr Robertson said: ‘We chose to study the Samsung Galaxy because it is a very popular phone which comes with working face recognition technology. However, we expect this technique to work across a wide range of phones and other automated recognition devices. It is very interesting that performance can be so much improved by copying a simple trick performed by the brain.”

About this technology research

Contact: David Robertson – University of York
Source: University of York press release
Image Source: The image is adapted from the University of York press release
Original Research: Full open access research for “Face Averages Enhance User Recognition for Smartphone Security” by David J. Robertson, Robin S. S. Kramer, and A. Mike Burton in PLOS ONE. Published online March 25 2015 doi:10.1038/ni.3132

Open Access Neuroscience Abstract

Face Averages Enhance User Recognition for Smartphone Security

Our recognition of familiar faces is excellent, and generalises across viewing conditions. However, unfamiliar face recognition is much poorer. For this reason, automatic face recognition systems might benefit from incorporating the advantages of familiarity. Here we put this to the test using the face verification system available on a popular smartphone (the Samsung Galaxy). In two experiments we tested the recognition performance of the smartphone when it was encoded with an individual’s ‘face-average’ – a representation derived from theories of human face perception. This technique significantly improved performance for both unconstrained celebrity images (Experiment 1) and for real faces (Experiment 2): users could unlock their phones more reliably when the device stored an average of the user’s face than when they stored a single image. This advantage was consistent across a wide variety of everyday viewing conditions. Furthermore, the benefit did not reduce the rejection of imposter faces. This benefit is brought about solely by consideration of suitable representations for automatic face recognition, and we argue that this is just as important as development of matching algorithms themselves. We propose that this representation could significantly improve recognition rates in everyday settings.

“Face Averages Enhance User Recognition for Smartphone Security” by David J. Robertson, Robin S. S. Kramer, and A. Mike Burton in PLOS ONE doi:10.1038/ni.3132.

Share this AI News
Join our Newsletter
I agree to have my personal information transferred to AWeber for Neuroscience Newsletter ( more information )
Sign up to receive our recent neuroscience headlines and summaries sent to your email once a day, totally free.
We hate spam and only use your email to contact you about newsletters. You can cancel your subscription any time.