Multi-modal biometrics combining multiple biometric types

Multi-modal biometrics use data from multiple biometrics, captured from the same user. This data is then combined or fused to improve the accuracy, security and usability of an identity solution. For example, a user might be asked to provide face and voice biometrics to authenticate themselves. Daon's biometric platforms gives organizations the reassurance of this multi-modal approach, for robust and accurate identification.

Multi modal screen

How do multi-modal biometrics work?

Typically, multiple biometrics are captured sequentially one after the other; the capture of subsequent biometrics can optionally be made conditional on the previous biometric authentication not reaching a sufficient level of security. Alternatively, it is possible to capture multiple biometrics simultaneously, by asking the user to speak a phrase while looking at the camera, for example.

What are the advantage of multi-modal biometrics?

Using two or more biometrics can help to compensate for poor quality data or faults in user positioning or capture sensors, improving overall accuracy and increasing a system's security.

Multi-modal systems also allow organizations to provide customers with a choice of biometrics depending on their preferences and current situation. Facial recognition might be a better choice on a noisy city street, while voice biometrics might be a more suitable to quieter locations though biometrics is rather robust.

What is multi-modal fusion?

Combining the information from two or more biometric modalities is called fusion. There are two different types of fusion – decision and score.

  • Decision Fusion: The results of each separate authentication must be confirmed e.g. the user must authenticate successfully with both face and voice.

Multi modal chart 2

With decision fusion, the strength or confidence level in each separate authentication is not taken into account. The user may have achieved a high matching score (99.99%) in one biometric but may have achieved a slightly lower than threshold in a second biometric (e.g. 1% confidence less than the pass score). In this case, the authentication attempt would fail.

  • Score fusion: This approach allows individual matching scores to be taken into account in the final fused score.

Multi modal chart 1

Strong authentication is one biometric can be used to balance out the confidence score achieved in another biometric to achieve an overall authentication. This could be done by taking an average of both scores, yielding a fused confidence percentage which is above a fused threshold. In this scenario, the combined results of both biometrics is better than either biometric on its own.

Find out more about how Daon’s IdentityX platform allows organizations to offer multiple biometrics and to combine biometrics, using fusion, for a multi-modal approach.