Facial Recognition Methods

Truora’s face validation system incorporates two main methods for facial recognition. The first and simplest one is similarity, which compares the reference picture with a selfie of the individual. The second and more sophisticated one is passive liveness, which compares the reference picture with a short non-guided video of the individual to detect if it pertains to a live person.

Similarity

The similarity method is designed to compare the facial features of a submitted selfie with those of a stored reference image. This process focuses solely on verifying that the two images belong to the same person by analyzing and matching specific facial characteristics.

Use Case: Typically used in scenarios where confirming the identity of the user is the primary concern. For instance, when a user logs into an account using facial recognition, the system compares the new selfie to the reference image stored during account creation.

Passive liveness

The passive liveness method goes beyond simple facial feature comparison. It involves comparing a reference image with a non-guided short video to ensure that the submitted media is from a live person and not a photograph or video replay. This method checks for subtle movements, changes in lighting, and reflections that are indicative of a live subject.

Benefits:

• There is no need to perform any kind of actions in the video, making the process seamless and user-friendly.
• Prevents print, replay, 3d masks and deepfake videos attacks.
• Increases the security of the identity verification.
• Keeps the verification process simple, secure and fast.

Use Case: Passive liveness is crucial in high-security environments where preventing spoofing attacks is essential. It ensures not only that the facial features match but also that the user is physically present at the time of verification.

Activating passive liveness and/or similarity

Currently, we have two parameters for activating face validation methods: subvalidations and liveness (soon to be deprecated). Please ensure to use only one of these parameters and note that liveness will be deprecated soon. If you are still using the old parameter and have questions after reading this guide, please Contact Support.

New Parameter

To set the subvalidations parameter:

In the Create Validation body request for face validation, use the following:

• Default value (similarity): Omit the subvalidations parameter or set it to similarity to perform face validation using similarity.
  • subvalidations: similarity
• Passive Liveness: Set the subvalidations parameter to passive_liveness to perform face validation using passive liveness.
  • subvalidations: passive_liveness
• Using both methods: You can combine both methods by setting subvalidations to include both similarity and passive_liveness.
  • subvalidations: passive_liveness
  • subvalidations: similarity

Old Parameter (soon to be deprecated)

To set the liveness parameter:

In the Create Validation body request for face validation, use the following:

• Default value (similarity): Omit the liveness parameter or set it to disabled to perform face validation using similarity.
  • liveness: disabled
• Passive Liveness and Similarity: Set the liveness parameter to passive to perform face validation using both passive liveness and similarity.
  • liveness: passive