Comment on page
Identity Check
Ensure that your clients are who they say they are.
Uses ComplyCube's Deep Learning and liveness detection, and facial recognition engine to analyze and calculate a similarity score between a provided selfie image and an ID document. This will determine if it’s the same person in both images.
It also leverages 3D face maps to determine whether a client has been previously enrolled using different details.
ComplyCube's liveness detection leverages state-of-the-art Presentation Attack Detection (PAD) technology which builds 3D facial maps, analyzes skin texture and micro-expressions, detects pixel tampering, recognizes masks and disguises, and employs many other proprietary techniques.
Our Identity Check can flag advanced spoofing vectors, including but not limited to:
- Printed Photo Attack: a fraudster uses a printed photo of an individual to present during verification.
- Printed Mask Attack: a fraudster uses a mask or a photo, often with cut-out holes, to blink, a common and crude test of liveness.
- Video Replay Attack: an imposter presents a video of an individual to fool a verification system that expects movement to establish liveness.
- 3D Mask Attack: a fraudster purchases or creates a silicon or plastic mask of an individual.
- Network Spoofing: a fraudster using a Virtual Private Network (VPN) to gain access to your platform.
In short, our platform does the heavy lifting to ensure that your customer is physically present and not an inanimate artifact, injected video, or a deepfake. For example, liveness detection will ensure that a person opening a bank account, making a purchase, or transferring money is present during the transaction and is not being impersonated through a stolen or fake identity.
Even though it is possible to run this check entirely through our API, we highly recommend that you run it through our Hosted Solution, Web SDK, or native Mobile SDKs. These SDKs guide your clients through the capture process, aiding them with completing the flow successfully with minimal friction and false-positive outcomes.
An Identity Check takes an average of 5 seconds to complete.
Last modified 2mo ago