FB pixel

Researchers improve smartphone face recognition by combining photos


University of York researchers have discovered that facial recognition features on smartphones can be drastically improved if users combined different images of themselves instead of a single ‘target’ image.

The research group, led by Dr. David Robertson, of the Department of Psychology’s FaceVar laboratory at York, found that combining several images to create an ‘average’ leads to a greater recognition rate across a range of daily settings.

The research paper, “Face Averages Enhance User Recognition for Smartphone Security“, published in the journal PLOS ONE, is based on several tests using the ‘face unlock’ system on Samsung Galaxy phones.

Researchers found that while the system was usually consistent in rejecting impostors, it often failed to recognize the actual smartphone owner.

However, when researchers morphed several different photos of the user to create an ‘average’, the performance of the facial recognition feature’ significantly improved, often to perfect levels..

The researchers’ technique is based on studies of human face recognition in which the brain forms abstract representations of the faces it recognizes, as well as the fact that humans are typically able to recognize their family and friends under various conditions.

‘We chose to study the Samsung Galaxy because it is a very popular phone which comes with working face recognition technology, said Dr. Robertson. “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.”

Article Topics

 |   |   |   | 

Latest Biometrics News


15 Replies to “Researchers improve smartphone face recognition by combining photos”

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Most Read This Week

Featured Company

Biometrics Insight, Opinion

Biometrics White Papers

Biometrics Events

Explaining Biometrics