Passive liveness

Proving liveness invisibly, from a single normal selfie.

The person takes an ordinary selfie, and the system decides in the background whether the face is a live human, with no prompt to blink or turn.

Liveness inference from intrinsic image and sensor properties with no user action required. Intrinsic properties are the qualities already present in the captured image, such as texture and lighting, rather than a response to a prompt.

A passive liveness check works like this:

1. The person takes a normal selfie or a short capture.
2. A model analyzes the image or frames for signs of a live face.
3. The system returns a live or not-live result.

The model looks for cues such as:

• Skin texture and fine detail that a print or screen does not reproduce.
• Reflections and lighting that differ between skin and glass or paper.
• Screen artifacts, such as the moire patterns that appear when a camera films a display.
• Depth cues and tiny natural movements, when the capture includes several frames.

Passive liveness sits next to terms that are easy to mix up:

• Active liveness asks the person to perform an action.
• Liveness detection is the umbrella term for both approaches.
• Conversion rate is the share of people who finish the flow, which passive checks tend to protect.