Facial Recognition Matching Level defines how strict the face comparison is when users authenticate with face recognition. Higher levels require a more accurate match and allow less face occlusion.
Available levels:
Low: Users can pass recognition if no more than 15% of the face is covered (mouth not included).
Normal: Users can pass recognition if no more than 10% of the face is covered (mouth not included).
High: Users can pass recognition if no more than 5% of the face is covered (mouth not included).
Highest: Users can pass recognition only when the face is almost completely visible.
Note
Higher matching levels improve security but may reduce recognition speed or tolerance in real deployment.
Face Living Recognition Matching Level defines the anti-spoofing strength of the device. It determines whether photos or screen images can be used to bypass face recognition.
Available levels:
Close: Paper photos and electronic photos can pass recognition.
Low: Black-and-white paper photos (single camera) and electronic photos (dual cameras) cannot pass recognition.
Normal: Black-and-white paper photos and electronic photos cannot pass recognition.
High: All paper and electronic photos (color or black-and-white) cannot pass recognition.
Highest: Only real human faces can pass recognition.
Note
Use higher liveness levels in high-security environments to prevent photo or screen spoofing attacks.