Spoof Score

In Liveness, liveness detection results are returned only in a quantitative value instead of a qualitative result.

Spoof score and Threshold

All APIs in Liveness return a spoofScore reflecting the probability of the proceeded face belonging to a spoof attack if the request is successful. The spoofScore is a float number ranging from 0.0 to 1.0.

Different image resolutions, angles, lighting conditions, etc., might affect the liveness score. The output is not necessarily the same, even if you send two identical images or videos.

A threshold is the standard to decide the whether the proceeded personnel is from a spoof attack based on the score returned in the API. We highly recommend determining your customized threshold value to compare with the spoof score and judge the qualitative result of the liveness detection. This logic should be implemented into your system.

Depending on the threshold, the result of liveness detection might be different. For example, suppose the threshold value is set to 0.50, and the spoof score is 0.40. The hypothesis of the the face being a spoof is rejected since 0.40 < 0.50. On the contrary, if the threshold is set to 0.30, we can accept the same hypothesis since 0.40 >= 0.30. The threshold setting is a trade-off between the false acceptance rate (FAR) and the false rejection rate (FRR). The higher the threshold, the more likely a false rejection and the less likely a false acceptance would happen.

Different businesses have different use scenes and different demands on liveness detection accuracy. Refer to each specific models for the recommended threshold value.