SensoryCloud AI - 支持Liveness的声纹生物特征识别

What is “Liveness”? In biometrics, Liveness Detection is a computer's ability to determine that it is interfacing with a physically present human being and not an inanimate spoof artifact or injected video/data.

Biometric data is the unique information that can be used to identify a person with accuracy. It includes uniquely identifiable features such as fingerprint, face recognition, iris, voice recognition. The increased acceptance of biometrics by consumers has encouraged the uptake of these systems on a wider scale.

“… biometric prints need not be kept secret, but the validation process must check for liveness of the readings.”

Liveness is a system ability to detect whether a biometric is from a live person or a spoof, so artefact or lifeless body part. By using AI systems in biometric detection it can help to make a much more secure system.

Liveness detection

Liveness detection is an anti-spoofing measure built into a biometric system, which is meant to confirm that the body part submitted for authentication is real.

Liveness detection is critical in presentation attacks where fraudsters try to break into a biometric system using fake fingerprints, masks, high definition pictures and so on. When it comes to voice biometrics, the most common spoofing attempt is by trying to manipulate voice biometrics using an audio recording (playback) of the victim. A more recent method to try and overcome voice biometrics is the use of synthetic speech tools.

How is it done?

Biometric system detection algorithms are not set up to detect ‘live’ from not live, they only need to match what is presented to the enrolled data and so granting or denying authentication. Liveness will not match but will look for other markers that ‘prove’ live presentation and so significantly reduce risks of a spoof attack.

There are 3 categories of liveness algorithms that can be used;

• Active Liveness detection. This requires a challenge and response, for example, the user will be prompted to make a facial action during a facial recognition capture, like a smile or blink. Users are fully aware of the liveness detection measure and so is less secure and poses more risk to fraud.
• Passive liveness detection. This will rely on background algorithms that assess artefacts in an image, such as an edge or skin texture as well as motion detection. This process does not require the user’s active participation and so is hidden to them so making it the most secure as it’s more difficult for fraudsters to attempt to circumvent it.
• Hybrid is also one that does not require user interaction but is not opaque and is observable by the fraudsters, making it potentially more vulnerable.

There are two common practices among fraudsters: playing a recording which matches the desired voiceprint, and using synthetic speech.

For more information about face biometrics liveness, goes to -

https://www.liveness.com/

A live demo of how SensoryCloud AI tackled the biometrics anti-spoofing issue by VoicePrint Recognition with LIVENESS. http://mpvideo.qpic.cn/0bc3z4aakaaabiapblkm7frfbt6daxhqabia.f10002.mp4?dis_k=877370ac450ac906e24fc721947d86dc&dis_t=1652763458&vid=wxv_2342410578373443585&format_id=10002&support_redirect=0&mmversion=false