We chose to use Liveness because it adds a certain level of trustworthiness to the data we consume
Chief Product Officer, Bunq
Add biometrics to your identity verification. See a real identity.
At Onfido, we take a new approach to identity—combining a person’s ID with their physical biometrics. Hundreds of our customers have added biometric verification to their document verification, with definitive results. Drivy (now Getaround) worked with us to include biometric verification in their flow, and saw a 22% decrease in theft attempts.
Protect against stolen IDs and impersonation.
Why is biometric verification so effective? Well, let’s say you’ve verified a document. You need Onfido’s biometric technology to verify that the document truly belongs to the person making the transaction. So biometrics adds a layer of protection against stolen IDs and impersonation attacks.
And it’s a simple task for your users. All they need to do is take a selfie or a short video. Using advanced facial scanning, Onfido then compares their facial biometrics to the photo on the ID, and generates a score based on the similarity of the faces.
Minimize customer friction. Catch stolen IDs.
It’s a simple step in your user experience. After your user submits the ID, Onfido guides the user to help them take the clearest selfie possible.
Onfido performs liveness checks to verify that the selfie is a live photo, and compares it to the photo in the ID. Seconds later, we calculate a score based on how similar the two faces are.
Not alike enough? We alert you to the possibility of impersonation fraud. But too alike? That could indicate a spoof attempt to pass off a photo of the ID as a selfie. We catch both in one step.
Use a selfie captured at sign-up later in your customer journey, too. Rather than two-factor authentication or a call center interaction, ask for a face with our Selfie Recheck service.
Onfido Selfie doesn’t just smooth the user journey. It also protects against a wide range of fraud attack vectors
So you’ll have the assurance you need that the person who owns the ID really is the person submitting it.
Uploading a photo from the web
Photo of the photo in the ID document
Photo of a printed photo
Photo of a digital screen
Uploading a digitally tampered selfie
Photo of a digitally tampered selfie
2D mask or a printout with eyeholes
Add the next layer of security: video.
Increased reward means increased incentive for fraud. For high value products, high risk users or highly regulated geographies, you need the next level of protection.
That’s where Onfido Video comes in. It deflects the same attack vectors as Selfie, with additional liveness coverage for new, sophisticated fraud techniques—from spoof selfies to deep fake videos, 2D to 3D latex masks.
Onfido prompts the user to take a short video on their phone, guiding them to perform a movement and read three randomly generated numbers aloud.
It only takes a few seconds. It’s just two simple actions. But behind the scenes, Onfido Video:
Combines advanced fraud defense techniques including motion tracking, lip sync and texture analysis to verify that the user is physically present.
Offers over 4,000 combinations of randomized actions, so fraudsters can’t cheat the system with pre-recorded videos.
Uses hybrid machine learning technology for maximum scalability, accuracy and speed.
Advanced algorithms comb through the video, assessing whether the user is there in person, and comparing their face to the photo on the ID.
Onfido Video isn’t just robust. It’s also user-friendly. It’s language agnostic, for the best global user experience. Plus, real-time face detection and head tracking provide instant user feedback—so you can see more successful recordings and higher pass rates.
How is Onfido’s Biometric Verification different?
Onfido’s API gives you maximum flexibility to match your risk appetite. You can decide on a per user basis which verification type to use. So, faced with a user from a high risk region, you might choose Video. But for lower risk transactions, you could specify Selfie. Minimize friction or maximize security—it’s all about your needs.
Onfido Video checks can outperform live-video calls in matching faces, detecting liveness, reducing bias and increasing consistency.
Plus, our machine learning models do not get tired or succumb to lapses in attention, compared with humans.
Onfido uses certified super recognizers. These are the 1% of humans that can match faces with an accuracy rate of 98%. The rest of us? We can only match 50%. Because they’re the best humans in the world at it, super recognizers make our tech—and your results—better. They review cases where our AI can’t return a conclusive result, and train our algorithms to make them even better.
Want to know more about identity verification? Read the most relevant and interesting updates, insights and research from our content hub.