Introducing Face Liveness Detection: Stop Spoofing Before It Starts

Identity verification is only as strong as its weakest point. You can have the best face-matching algorithm in the world, but if someone holds up a photo of another person's face to the camera, none of that matters. Liveness detection closes that gap by determining whether a face is physically present during capture, or whether it is a spoofing attempt.

Sightengine's Face Liveness Detection API uses passive liveness detection: it analyzes a single photo, with no user interaction required. No head turns, no blinks, no eye tracking. The user simply presents their face to the camera, and the API determines whether the face is live or spoofed.

What it detects

The API identifies a wide range of presentation attacks:

Replay attacks (face on a screen)

Digital reproductions on phones, tablets, or monitors. This includes cases where the edge of the screen or monitor is not visible.

Face on a printed paper

Physical prints, photos, magazine covers, posters, etc.

Face added as postprocessing

Digital overlays or elements inserted after the image was captured.

2D face masks

Physical masks that are flat and two-dimensional.

3D masks

3D attempts using materials such as silcone, paper, or cardboard.

Cutouts

Physical cutouts of a full body, of an upper body, or of a face.

Face cutouts

Physical cutouts of a face.

Deepfakes

AI-generated face swaps or visual manipulations.

GenAI images

Entirely synthetic faces generated by AI models.

This breadth of coverage matters because attackers do not limit themselves to one technique. A system that catches screen replays but misses printed photos leaves a door open.

How it works

The workflow is straightforward:

1. Capture an image. The user takes a selfie or presents their face to a camera. A single static photo is all that's needed.
2. Send it to the API. Submit the image to Sightengine's check.json endpoint with the liveness model.
3. Get the result. The API returns a liveness score and, if the image is rejected, a clear rejection reason.

Because the detection is passive, there is no friction for the end user. No multi-step challenges, no waiting for prompts. This makes it suitable for high-volume onboarding flows where drop-off rates matter.

Quality validation built in

Before performing liveness analysis, the API validates that the image meets the requirements for a reliable assessment:

• A single dominant face must be present (background faces are acceptable if significantly smaller).
• The face must occupy between 1% and 95% of the image area.
• The face must be clearly visible, adequately lit, and unobstructed.

Images that do not meet these criteria are rejected with a specific reason, so your application can prompt the user to retake the photo rather than returning an unreliable result.

Use cases

Liveness detection is a critical building block for any workflow that relies on verifying a person's identity through their face:

• Identity verification and KYC. Financial services, fintech, and regulated industries need to confirm that the person opening an account is physically present, not someone holding up a stolen ID photo.
• User onboarding. Platforms that require selfie verification during sign-up can use liveness detection to prevent fake account creation at scale.
• Access control. Any system using facial recognition for access (physical or digital) benefits from liveness checks to prevent bypass through photos or videos of authorized users.
• Fraud prevention. Insurance claims, loan applications, marketplace transactions: anywhere a face photo is used as proof of identity, liveness detection adds a layer of assurance.

Get started

Face Liveness Detection is available now through the Sightengine API. It works with the same API keys and endpoint you already use for other Sightengine models, so integration is straightforward.

• Read the full documentation for detailed integration instructions and code examples.
• Already using Sightengine? Just add liveness to your models parameter to start checking images immediately.

Read more

Sightengine Ranks #1 in AI-Media Detection Accuracy: Insights from an Independent Benchmark

Learn how Sightengine performed in an independent AI-media detection benchmark, outperforming competitors with advanced methodologies.

Sexual abuse and unsolicited sex — Content Moderation

This is a guide to detecting, moderating and handling sexual abuse and unsolicited sex in texts and images.

Head back to the Knowledge Center