Detect Facial Recognition Bypass Using Deep Fakes in Android Apps

Last updated May 15, 2023 by Appdome

Learn to Detect Deep Fake in Android apps, in mobile CI/CD with a Data-Driven DevSecOps™ build system.

What is Deep Fake?

Mobile app deep fake facial ID bypass uses digital manipulation techniques to create a fake facial image that can bypass facial authentication on a mobile device or within a specific mobile application. Attackers can use this technique to gain unauthorized access to a device, such as a smartphone or tablet, by fooling the device’s facial recognition technology.

There are many mechanisms used for spoofing images, such as intercepting and modifying the camera preview data from other apps while the preview shown to the user remains unchanged. In most cases, this mechanism is used for stealing sensitive data such as bank account details, bank checks, or passwordsץ

 

How does Appdome Detect Deep Fake?

Upon detecting the operation of the deep fake mechanism, regardless of the application that uses the Android camera, Appdome will close the mobile app and display a user notification message of the detection.

Detecting Deep Fake on Android apps by Using Appdome

On Appdome, follow these 3 simple steps to create self-defending Android Apps that Detect Deep Fake without an SDK or gateway:

  1. Upload the Mobile App to Appdome.

    1. Upload an app to Appdome’s Mobile App Security Build System

    2. Upload Method: Appdome Console or DEV-API
    3. Android Formats: .apk or .aab
  2. Build the feature: Deep Fake Detection.

    1. Building Deep Fake Detection by using Appdome’s DEV-API:

      1. Create and name the Fusion Set (security template) that will contain the Deep Fake Detection feature as shown below:
      2. fusion set that contains Deep Fake Detection

        Figure 1: Fusion Set that will contain the Deep Fake Detection feature
        Note: Naming the Fusion Set to correspond to the protection(s) selected is for illustration purposes only (not required).

      3. Follow the steps in Sections 2.2.1-2.2.2 of this article, Building the Deep Fake Detection feature via Appdome Console, to add the Deep Fake Detection feature to this Fusion Set.

      4. Open the Fusion Set Detail Summary by clicking the “...” symbol on the far-right corner of the Fusion Set, as shown in Figure 1 above, and get the Fusion Set ID from the Fusion Set Detail Summary (as shown below): fusion Set Detail Summary image

        Figure 2: Fusion Set Detail Summary
        Note: Annotating the Fusion Set to identify the protection(s) selected is optional only (not mandatory).

      5. Follow the instructions below to use the Fusion Set ID inside any standard mobile DevOps or CI/CD toolkit like Bitrise, App Center, Jenkins, Travis, Team City, Cirlce CI or other system:
        1. Build an API for the app – for instructions, see the tasks under Appdome API Reference Guide
        2. Look for sample APIs in Appdome’s GitHub Repository
    2. Building the Deep Fake Detection feature via Appdome Console

      To build the Deep Fake Detection protection by using Appdome Console, follow the instructions below.

      1. Where: Inside the Appdome Console, go to Build > Anti Fraud Tab > Mobile Malware Prevention section
      2. How: Toggle (turn ON) Deep Fake Detection, as shown below.
        If needed, Customize the Threat Notification to be displayed to the mobile end-user in a standard OS dialog notification when Appdome Detects Deep Fake. Deep Fake Detection option

        Figure 3: Detect Deep Fake option
        Note: The App Compromise Notification contains an easy to follow default remediation path for the mobile app end user. You can customize this message as required to achieve brand specific support, workflow or other messaging.

      3. When you select the Deep Fake Detection you'll notice that your Fusion Set you created in step 2.1.1 now bears the icon of the protection category that contains Deep Fake Detection

        Fusion Set applied Deep Fake Detection

        Figure 4: Fusion Set that displays the newly added Deep Fake Detection protection

      4. Select the Threat-Event™ in-app mobile Threat Defense and Intelligence policy for Deep Fake Detection:
        1. Threat-Events™ OFF > In-App Defense

          If the Threat-Events™ setting is cleared (not selected). Appdome will detect and defend the user and app by enforcing Detect Deep Fake.

        2. Threat-Events™ ON > In-App Detection

          When this setting is used, Appdome detects a hook on a camera-related Java method and passes Appdome’s Threat-Event™ attack intelligence to the app’s business logic for processing, enforcement, and user notification. For more information on consuming and using Appdome Threat-Events™ in the app, see section Using Threat-Events™ for Deep Fake Detection Intelligence and Control in Mobile Apps.

        3. Threat-Events™ ON > In-App Defense

          When this setting is used, Appdome detects and defends against Deep Fake (same as Appdome Enforce) and passes Appdome’s Threat-Event™ attack intelligence the app’s business logic for processing. For more information on consuming and using Appdome Threat-Events™ in the app, see section Using Threat-Events™ for Deep Fake Detection Intelligence and Control in Mobile Apps.

      5. Click Build My App at the bottom of the Build Workflow (shown in Figure 3).
    Congratulations!  The Deep Fake Detection protection is now added to the mobile app
  3. Certify the Deep Fake Detection feature in Android Apps.

    After building Deep Fake Detection, Appdome generates a Certified Secure™ certificate to guarantee that the Deep Fake Detection protection has been added and is protecting the app. To verify that the Deep Fake Detection protection has been added to the mobile app, locate the protection in the Certified Secure™ certificate as shown below: Deep Fake Detection shown in Certificate secure

    Figure 5: Certified Secure™ certificate

    Each Certified Secure™ certificate provides DevOps and DevSecOps organizations the entire workflow summary, audit trail of each build, and proof of protection that Deep Fake Detection has been added to each Android app. Certified Secure provides instant and in-line DevSecOps compliance certification that Deep Fake Detection and other mobile app security features are in each build of the mobile app

Using Threat-Events™ for Deep Fake Intelligence and Control in Android Apps

Appdome Threat-Events™ provides consumable in-app mobile app attack intelligence and defense control when Deep Fake is detected. To consume and use Threat-Events™ for Deep Fake in Android Apps, use registerReceiver in the Application OnCreate, and the code samples for Threat-Events™ for Deep Fake shown below.

The specifications and options for Threat-Events™ for Deep Fake are:

In-App Defense

In-App Detection

Appdome detected hook on Java method related to camera.

Threat-Event™ Elements Detect Deep Fake Method Detail
Appdome Feature Name Deep Fake Detection
Threat-Event Mode
OFF, IN-APP DEFENSE Appdome detects, defends and notifies user (standard OS dialog) using customizable messaging.
ON, IN-APP DETECTION Appdome detects the attack or threat and passes the event in a standard format to the app for processing (app chooses how and when to enforce).
ON, IN-APP DEFENSE Uses Appdome Enforce mode for any attack or threat and passes the event in a standard format to the app for processing (gather intel on attacks and threats without losing any protection).
Certified Secure™ Threat Event Check
Visible in ThreatScope™
Developer Parameters for Detecting Deep Fake Threat-Event™
Threat-Event NAME FaceIDBypassDetected
Threat-Event DATA reasonData
Threat-Event CODE reasonCode
Threat-Event REF 6939
Threat-Event SCORE
currentThreatEventScore Current Threat-Event score
threatEventsScore Total Threat-events score
Threat-Event Context Keys
message Message displayed for the user on event
externalID The external ID of the event which can be listened via Threat Events
osVersion OS version of the current device
deviceModel Current device model
deviceManufacturer The manufacturer of the current device
fusedAppToken The task ID of the Appdome fusion of the currently running app
kernelInfo Info about the kernel: system name, node name, release, version and machine.
carrierPlmn PLMN of the device
deviceID Current device ID
reasonCode Reason code of the occured event
buildDate Appdome fusion date of the current application
devicePlatform OS name of the current device
carrierName Carrier name of the current device
updatedOSVersion Is the OS version up to date
deviceBrand Brand of the device
deviceBoard Board of the device
buildUser Build user
buildHost Build host
sdkVersion Sdk version
timeZone Time zone
deviceFaceDown Is the device face down
locationLong Location long
locationLat Location lat
locationState Location state
wifiSsid Wifi SSID
wifiSsidPermissionStatus Wifi SSID permission status

With Threat-Events™ enabled (turned ON), Android developers can get detailed attack intelligence and granular defense control in Android applications and create amazing user experiences for all mobile end users when Deep Fake is detected.

The following is a code sample for native Android apps, which uses all values in the specification above for Deep Fake Detection:

Prerequisites to Using Deep Fake Detection:

To use Appdome’s mobile app security build system to Detect Deep Fake , you’ll need:

Using Appdome, there are no development or coding prerequisites to build secured Android Apps by using Deep Fake Detection. There is no SDK and no library to code or implement in the app and no gateway to deploy in your network. All protections are built into each app and the resulting app is self-defending and self-protecting.

Releasing and Publishing Mobile Apps with Deep Fake Detection

After successfully securing your app by using Appdome, there are several available options to complete your project, depending on your app lifecycle or workflow. These include:

All apps protected by Appdome are fully compatible with any public app store, including Apple App Store, Google Play, Huawei App Gallery and more.

Features Similar to Deep Fake Detection

Here are a few related resources:

If you have any questions, please send them our way at support.appdome.com or via the chat window on the Appdome platform.

Thank you!

Thanks for visiting Appdome! Our mission is to secure every app on the planet by making mobile app security easy. We hope we’re living up to the mission with your project.

NEED HELP?

let's solve it together

ScottMaking your security project a success!
By filling out this form, you opt-in to recieve emails from us.