td_liveliness_detection 1.0.1

Face liveness detection for Flutter with blink, smile and head-turn challenges, an arch-shaped camera frame and selfie capture, built on Google ML Kit.

td_liveliness_detection

Face liveness detection for Flutter. The user is asked to perform simple challenges (blink, smile, turn head) in front of the front camera, and a selfie is returned once every challenge passes. Detection runs on-device with Google ML Kit, so no paid API or network call is needed.

Features

• Blink, smile, and left/right head-turn challenges, asked in random order
• Waits for a neutral face between challenges, so one expression can't pass two steps
• Arch-shaped camera frame with a ring that changes colour as the user progresses
• Step indicator, live instructions below the frame, and a success animation
• Returns the captured selfie as an XFile
• Android and iOS

Installation

dependencies:
  td_liveliness_detection: ^1.0.0

Android

In android/app/src/main/AndroidManifest.xml:

<uses-permission android:name="android.permission.CAMERA" />

Set minSdk to at least 21 in android/app/build.gradle(.kts) (ML Kit requirement).

iOS setup

1. Minimum iOS 16.0. In ios/Podfile:

   platform :ios, '16.0'
   

   Also set iOS Deployment Target to 16.0 for the Runner target in Xcode.

2. Camera usage description in ios/Runner/Info.plist:

   <key>NSCameraUsageDescription</key>
   <string>The camera is used to confirm you are a real person.</string>
   

3. If your app uses permission_handler, enable the camera permission in the post_install block of ios/Podfile:

   post_install do |installer|
     installer.pods_project.targets.each do |target|
       flutter_additional_ios_build_settings(target)
       target.build_configurations.each do |config|
         config.build_settings['GCC_PREPROCESSOR_DEFINITIONS'] ||= [
           '$(inherited)',
           'PERMISSION_CAMERA=1',
         ]
       end
     end
   end
   

4. Test on a real iPhone. The iOS Simulator has no front camera.

Usage

import 'package:camera/camera.dart';
import 'package:td_liveliness_detection/td_liveliness_detection.dart';

final XFile? selfie = await FlutterLivenessDetection().start(
  context: context,
  expressions: [
    Expression.eyeblink,
    Expression.smile,
    // Expression.leftPose,
    // Expression.rightPose,
  ],
);

if (selfie != null) {
  // Liveness passed – use selfie.path
} else {
  // User cancelled
}

Request camera permission before calling start (for example with permission_handler). See the example/ app for a complete flow.

Expressions

Expression	Passes when
Expression.eyeblink	Either eye is closed (open probability < 0.3)
Expression.smile	Smiling probability > 0.5
Expression.leftPose	Head turned left (yaw > 10°)
Expression.rightPose	Head turned right (yaw < -10°)

License

MIT