DEV Community

HarmonyOS
HarmonyOS

Posted on

Face Liveness Detection in HarmonyOS: A VisionKit Implementation Guide

Read the original article：Face Liveness Detection in HarmonyOS: A VisionKit Implementation Guide

Introduction

Hello everyone! In this article, I’ll talk about Huawei’s HarmonyOS NEXT. It comes with VisionKit, a powerful tool that lets developers use AI-powered visual features. Here, I’ll focus on the Face Liveness Detection (interactiveLiveness) feature and explain how to integrate it into an app with examples.

📌 What is Face Liveness Detection?

Face Liveness Detection is a technology used to distinguish real human faces from fake ones (like photos, videos, or masks) through a device’s camera. In HarmonyOS NEXT, this feature verifies authenticity by asking users to perform interactive actions, such as blinking or nodding their head.

🎯 Use Cases

✅ Attendance tracking systems
✅ Secure app logins (identity verification)
✅ Fake user detection
✅ ID verification support

🧑‍💻 Development Steps

1. Add Required Module & Permissions

import { interactiveLiveness } from '@kit.VisionKit';
import { common, abilityAccessCtrl, Permissions } from '@kit.AbilityKit';
Enter fullscreen mode Exit fullscreen mode

Add camera permission to your module.json5 file.

"requestPermissions":[
  {
    "name": "ohos.permission.CAMERA",
    "reason": "$string:camera_desc",
    "usedScene": {"abilities": []}
  }
  ]
Enter fullscreen mode Exit fullscreen mode

2. Design the User Interface

@State routeMode: string = "replace";
@State actionsNum: number = 3;

Radio({ value: "replace", group: "routeMode" }).checked(true)
Radio({ value: "back", group: "routeMode" }).checked(false)

TextInput({ placeholder: "Enter 3 or 4" }).type(InputType.Number)
Enter fullscreen mode Exit fullscreen mode

3. Add a Start Button

Button("Start detection")
  .onClick(() => {
    this.startDetection();
  })
Enter fullscreen mode Exit fullscreen mode

4. Handle Permissions & Setup

private startDetection() {
  abilityAccessCtrl.createAtManager().requestPermissionsFromUser(this.context, this.array).then((res) => {
    if (res.permissions[0] === "ohos.permission.CAMERA" && res.authResults[0] === 0) {
      this.runLivenessDetection();
    }
  })
}
Enter fullscreen mode Exit fullscreen mode

5. Start Liveness Detection

private runLivenessDetection() {
  let config = {
    isSilentMode: "INTERACTIVE_MODE",
    routeMode: this.routeMode,
    actionsNum: this.actionsNum
  }

  interactiveLiveness.startLivenessDetection(config).then((success) => {
    if (success) {
      this.getDetectionResult();
    }
  }).catch((err) => {
    console.error("Detection failed", err);
  });
}
Enter fullscreen mode Exit fullscreen mode

6. Get the Result

private getDetectionResult() {
  interactiveLiveness.getInteractiveLivenessResult().then(result => {
    console.log("Liveness type:", result.livenessType);
  }).catch(err => {
    console.error("Failed to get result:", err);
  });
}
Enter fullscreen mode Exit fullscreen mode

🛠️ About the APIs

The interactiveLiveness module provides the following key functionalities for implementing face liveness detection:

⚙️startLivenessDetection(config: InteractiveLivenessConfig): Promise
Launches the face liveness detection screen using the provided configuration. The result of the redirection (success or failure) is returned as a Promise.

⚙️getInteractiveLivenessResult(): Promise
Retrieves the result of the liveness detection asynchronously. The returned result includes the detection mode, the best image captured, and optionally a secured buffer and certificate chain (for image security verification scenarios).

⚙️InteractiveLivenessConfig
A configuration object that allows you to customize the liveness detection behavior. It includes parameters such as:

  • isSilentMode: Defines the detection mode. Only INTERACTIVE_MODE is currently supported.
  • actionsNum: Specifies how many user actions (like blinking, nodding, etc.) will be randomly selected for detection. Only 3 or 4 actions are supported.
  • routeMode: Determines how the app navigates after detection, using either REPLACE_MODE or BACK_MODE.
  • successfulRouteUrl / failedRouteUrl: Optional paths to navigate after success or failure.
  • Additional options such as challenge (for security scenarios), speechSwitch (voice guidance), and isPrivacyMode (requires permission).

⚙️Enum Types for Configuration:

  • DetectionMode: Defines the detection method. Currently, only INTERACTIVE_MODE is available.
  • ActionsNumber: Specifies the number of required actions (3 or 4).
  • RouteRedirectionMode: Controls routing behavior after detection (BACK_MODE or REPLACE_MODE).

📌 For full details, check the official API documentation.

🎉 Conclusion

The interactiveLiveness feature offered by VisionKit provides significant advantages for user verification and preventing fake face attacks in your HarmonyOS applications. It’s easy to use, runs on-device, and responds very quickly according to device performance.

If you’re developing a security-focused application or just want to add a user verification layer, this feature is perfect for you.

References

Face Liveness Detection

Written by Baris Tuzemen

Top comments (0)