iOS users pass facial biometric checks faster than applicants on Windows

Otherwise known as “face authentication,” facial biometric checks are designed to protect online businesses from fraud by ensuring that users are truly present during remote verification. Face authentication is also a great step towards global digital equality as it lets honest users easily enter the online services they need regardless of their gender, nationality, income or digital skills.
Based on data collected globally in 2021, Sumsub analyzed the performance of facial biometric checks and differences across countries and devices, along with common failures and manipulation attempts. Here are the most interesting facts:

•
•
• The average global pass rate for facial biometric checks is currently
• 95%,
• with some differentiation between countries. For comparison, ‘selfie with ID’ checks have an average pass rate of around
• 83%.
• This suggests that it’s more convenient for people to get authenticated by showing their face, rather than taking a selfie while holding their ID card. Though facial biometric checks can be used alone, they’re usually combined with ID document verification,
• which effectively replaces ‘selfie with ID’ checks.
•

•
•
• In the UK, Japan, Canada, Bahrain, Hong Kong, Singapore and the EU, liveness checks usually take
• less than 35 seconds. In the US, they take about 45 seconds—whereas in Nigeria, Venezuela, Benin, Angola and Israel the average duration is 1.5 minutes. These disparities could be due to the check’s required facial movements being not as habitual for users
• in certain countries.
•

•
•
• There are striking differences in the average amount of time needed to complete facial biometric
• checks for users on different operating systems. For Linux, Windows and Android users, the process takes more than 50 seconds; whereas applicants on MacOS and iOS (Apple) devices need less than 39 seconds. Again, this difference could be a sign that Apple
• users are more accustomed to face authentication.
•

•
•
• 94% of failed facial biometric checks are due to improper conditions. Almost half of these (47%)
• are due to poor video quality or failure to perform the movements required to complete the check. There are also situations when the user’s face is too small (27%) or too large in the frame (16%). In 5% of cases, the applicant’s face is otherwise not properly
• visible.
•

•
•
• Possible digital manipulation accounts for almost 6% of all errors occurring during facial biometric
• checks. Of these, the most common (70%) are bypass attempts such as screenshots, photo manipulation, physical/digital masks, or sophisticated deepfakes and 3D models. Other common attempts at defeating the face authentication system involve the use of different
• devices or people (16%). The remaining cases (<14%) have to do with forced verification and malevolent attempts with devices placed too close to the applicant’s face.