Sightcorp Liveness Detection
Spoof attacks, meaning impersonation attacks where a representation of a genuine user’s face (e.g. video, photo) is used by another person to gain access to a secure (digital) space is one of the most common and worrisome challenges for Facial Recognition systems. At Sightcorp we are dedicated to building software that is not only accurate and stable but also safe. Therefore, we have developed and incorporated a ‘liveness check’ into our software so that you as an app operator can be certain that the person you are verifying is alive and not just a capture of that person in the form of an image, print-out, a screen, a video or a mask.
Within the industry, there are different approaches to tackle this issue. The current solutions landscape for liveness detection systems is broadly divided into:
- Active liveness detection. Such systems require the user to follow a random set of instructions. The systems tracks whether the user followed the instructions correctly to judge whether it was a real user or a pre-recorded/static image. While being arguably the most robust, these systems do require active participation from the user for authentication.
- Passive liveness detection. Such systems judge whether the user is real or not without any extra effort from the user itself. While offering a much simpler user experience, real world foolproof implementation of such systems is hard to build. There is no doubt however that with time such systems will be widely used as technology catches up and solves most of the current issues.
- Hybrid. Such systems use ‘creative’ approaches to tackle some of the issues of passive systems while not needing as much effort from the user as for a typical active check system.
Sightcorp’s active Liveness Detection is designed to help FaceMatch SDK clients detect & confirm the liveliness of their end users to rule out any spoof attacks. SDK customers are able to configure their own liveness detection check with a new sample which is now included by default in the FaceMatch SDK. The number of actions to be checked is configurable, the sequence of challenges is completely random and for extra safety, the user can be asked to perform multiple random gestures in a sequence.
How does it work?
The liveness check asks users to perform a random sequence of motions with their head (e.g. move up, move down) and asks them to smile on command. The FaceMatch software verifies if the sequence has been performed properly by the end user before a user is certified as “alive”. In case the liveness check is not passed, the user will be provided with feedback to explain which motions didn’t match the instructions.
How will it work in the future?
Sightcorp is continuously working on developing and improving its technology and algorithms to provide customers with more secure features and users with a more seamless experience. While our current liveness detection requires active user participation, we envision offering a solution that determines aliveness passively in the near future.
In addition to the SDK, liveness detection will soon be available as a standard function in our FaceMatch SDK as a Service as well. This will enable even more users to use this piece of essential technology in their applications.
Want to know more? Contact our sales team to find out how Liveness Detection can be the right solution to confirm and detect liveness of your users!
Try it out today!