Confirm the person. Keep no biometric data.

What AcuFace biometric facial matching does

Key capabilities for adding this module to a document or identity workflow.

• 1:1 facial comparison and liveness

  Compares a live capture with the portrait on the document, and checks the capture is live.

• No biometric data stored

  No templates, images or embeddings are retained.

• Advanced liveness detection

  Designed to detect real people and block spoofing attempts using photos, videos, masks and deepfakes.

• On-premise or cloud

  AcuFace can run on your own servers as well as in the Acupixel cloud.

  AcuFace only

• REST APIs, SDKs and webhooks

  Integrate AcuFace into an existing product and return its result to the workflow your application controls.

• Evaluated in NIST FRVT 1:1

  The AcuFace engine has been evaluated in NIST FRVT 1:1 Face Verification.

Privacy by design, from comparison to result

AcuFace is built to answer whether two faces match and whether the presented person is live, without building a store of biometric data.

• No biometric data stored

  AcuFace does not retain biometric templates, facial images or embeddings after the comparison.

• Mathematical comparison

  The proprietary Face-Int algorithm creates and compares mathematical representations for the transaction rather than storing biometric templates.

• 1:1 comparison

  Compare the live person with the portrait from the identity document or another reference image.

• Deployment under your control

  Run AcuFace on-premise or use the Acupixel cloud, depending on the security and infrastructure model your organization needs.

Liveness built to challenge common spoofing attempts

Advanced liveness detection is designed to confirm a real person is present and block presentation and synthetic-media attacks.

• Photo attacks

  Detect attempts to present a printed or displayed photograph in place of a live person.

• Video replays

  Challenge attempts to replay recorded facial footage to the capture experience.

• Masks

  Identify mask-based spoofing attempts intended to imitate the reference identity.

• Deepfakes

  Add liveness protection designed to block synthetic and manipulated facial media.

From two images to one usable result

Use the capture experience and reference source that fit your application, then return the result to the workflow you already control.

1. Provide the reference portrait

   Use the portrait from an identity document or another approved reference image.

2. Capture the live person

   Submit the live facial capture and run the liveness check through your cloud or on-premise AcuFace deployment.

3. Receive the result

   Return a 1:1 match score and liveness determination through a REST API, SDK or webhook integration.

NIST wording describes an evaluation. It does not imply certification, approval, ranking or a particular performance score.

Related products

Combine focused Acupixel modules to build the verification workflow your product needs.

• Document forensic authentication

  Evaluate whether a presented identity document appears genuine, altered, copied, recaptured, inconsistent or otherwise suspicious. APX-DFA combines document-template analysis, OCR, barcode and MRZ cross-validation, and supported forensic imaging to return clear authentication findings.

  Input Images of supported driver licences, ID cards, passports and other supported identity documents.
  Output OCR and machine-readable data plus authentication findings, supporting checks and a Pass, Warning or Fail-style result.

• AcuCapture

  Browser-based remote capture, with no app download for the person presenting the document.

  Input A one-time link sent to a person's phone.
  Output Their document and face, captured remotely and returned to your business.

Related documentation

Review the guides and endpoint details for planning an implementation.

• Process Facial Recognition And Liveness From Files

  The purpose of this endpoint is to perform a comparison between two images - typically a selfie and an image taken from an identity document. The endpoint verifies whether the person in both images is the same and allows performing a liveness check on the selfie image.