Applet - User Guide

Effective Date: 2026-01-28

Overview

This app is part of a testing program designed to test how well the app can distinguish between real human faces and fake faces (such as photos or videos). The purpose of this testing program is to collect data of real and fake human faces under various lighting conditions to support the development of machine learning models for facial liveness detection.

Please see below for a step-by-step guide on how to use the app (instructions are also included in the applet):

Before You Begin

Please make sure you:

Have a smartphone device with a working front-facing/selfie camera.

Have access to fake faces (for example, printed images or faces displayed on another screen/monitor).

Step-by-Step Instructions

Preparation

Please gather a handful of fake faces that can be found in your home:

This ensures the app is tested with a variety of faces.

You will need consent from the people in the pictures/videos if you are to use them as fake faces.

For example, make sure ‘Uncle Bob’ in the family photograph is okay with you using his picture as a ‘fake face’ (royalties not included*).

Lighting Setup

After you have organized your fake faces, please find an appropriate lighting environment for the selected applet test.

Tap ‘Start Scenario’ under one of the tests and follow the on-screen instructions to confirm your lighting condition and proceed with testing.

Face Positioning

Please position faces in the camera view:

First, position both real faces in the center and clearly visible.

Once both real faces are detected, add the fake face as instructed.

Head Movements

Next, you’ll be asked to perform a circular head movement.

When both real and fake faces are detected, slowly move each real face in a circular motion. The fake face does not need to move.

Labeling the Faces

In the final step, you will be asked to label each face.

When prompted, please select the option that best describes each face.