Ver-ID FAQ

Q: How accurate is face recognition?

A: Applied Recognition’s Ver-ID products all utilize the company’s FaceLocate ™ algorithm, trained for over a decade to detect and recognize hundreds of thousands of faces. FaceLocate can achieve up to 99.99% accuracy. An application can be tuned to correctly authenticate a user against their registered faceprint 99.99% of the time. To better understand accuracy ratings and related factors read our blog post here: http://appliedrecognition.com/blog/accuracy

Q: What impacts face recognition accuracy?

There are a number of variables that can impact the accuracy of face recognition including facial hair and make-up, and whether the user is wearing glasses or a hat. But by far the biggest accuracy challenge is illumination. Bright lights in the background or insufficient light can impact the ability of the camera to take a picture that can be matched against the stored face print.

Actual changes in the user’s appearance from aging, tanning or weight loss or gain can also impact accuracy, but to a lesser extent than illumination.

Q: What anti-spoofing methods are available?

Ver-ID uses state of the art anti-spoofing methods including liveness detection and multiple poses. These have been proven to be the most effective anti-spoofing options. For more information on anti-spoofing read our blog: http://appliedrecognition.com/blog/winks-blinks-smiles-and-poses-which-is-better-at-anti-spoofing

Q: How do anti-spoofing features affect the user experience?

A: Ver-ID’s anti-spoofing protections are invisible to the user. Once the user registers their faceprint using multiple poses, Ver-ID uses the data from the multiple poses during authentication, but the process takes place in milliseconds as part of the user login, so although anti-spoofing measures are taking place, the user doesn’t experience any inconvenience.

Q: Can Ver-ID be fooled with a photo or video?

A: Ver-ID is able to tell the difference between a live face and a photo 99.5% of the time. No biometric authentication solution is 100% accurate.

Q: How does face recognition compare to fingerprint or retina biometric authentication?

A: Retina scans take about 15 seconds of concentrated stare, whereas face recognition happens in a fraction of a second which makes face recognition more convenient. Face recognition can be used in conjunction with a registered photo ID providing an additional level of security and authentication. And face recognition only requires a common device camera, found on millions of devices including tablets, PCs, phones, ATM’s, security systems, etc; fingerprint authentication requires a fingerprint scanner, not found on the majority of smartphones.

Q: How can I incorporate a government or corporation-issued photo ID for additional security?

A: There are two ways Ver-ID can be used along with a photo ID for added security:

  1. A photo ID can be scanned and captured along with a real-time face authentication adding visible proof of identity. In this case the technology is confirming that the person in the faceprint matches the person in the photo ID.
  2. A faceprint can be matched against a stored photo ID authenticating the user against the stored photo providing digital proof of identity. In this case the authentication system is confirming the faceprint matches the photo ID.

Q: How does a user register their faceprint?

A: The registration process takes just a couple of minutes. The user opens a Ver-ID-enabled app on their mobile device. The app instructs them to hold a few different poses, taking pictures of each pose. Ver-ID maps 60+ points on the face, in each pose, and creates a unique identification code for that user. All of the information is stored on the user’s device ensuring privacy and security.

 

Q: What devices are supported?

A: Ver-ID Authentication, Ver-ID Credentials and Ver-ID Transactions run on all current iOS and Android devices. Ver-ID for Windows runs on Windows 7 and higher.

 

Q: What specific devices have you tested?

A: We test on a cross section of common devices, but it is impossible to test on all manufacturers and models.

 

Q: What are device requirements?

  • iOS ARM v7, ARM v7S, ARM 64. Ver-ID SDK can be added as a framework to any iOS project (9.1+)
  • Android (ARM) – Android projects targeted at API level 14+ can import the Ver-ID SDK from our Maven repository.

Q: How do I add Ver-ID to my application?

A: Adding Ver-ID to your application takes just a few lines of code. Ver-ID evaluation SDKs are available upon request at www.AppliedRecognition.com.

Q: Is there technical documentation available?

A: Our SDKs include technical documentation and our team of engineers is available for support. Visit www.AppliedRecognition.com to request an evaluation SDK.

Q: What standards do you support?

A: Ver-ID is compliant with Anti-Money Laundering (AML) and Know Your Customer (KYC) requirements, opening the door to digital client on boarding for the financial services sector.

Ver-ID Transactions is compliant with electronic signature legislation enacted by countries around the world. An ideal e-signature will create evidence equivalent to paper and ink. Ver-ID’s face and photo ID authenticated credentials make electronic signature stronger than paper and ink.

Q: What is Ver-ID current pricing?

A: Our pricing varies depending on payment model and volume. Please contact Sales@AppliedRec.com to discuss your solution needs.

If you have other questions please contact us at Sales@AppliedRec.com