face_live #

A Flutter plugin for real-time face liveness detection using native camera and ML Kit face detection.

Features #

• High-performance native camera preview (CameraX on Android, AVFoundation on iOS)
• Real-time ML Kit face detection with head movement tracking
• Progress updates (0-100%) based on head yaw rotation coverage
• Video recording of the liveness session for verification
• Cross-platform support for Android and iOS

Getting started #

Add this to your package's pubspec.yaml file:

dependencies:
  face_live: ^0.2.0

Platform Setup #

Android #

Add the following permissions to your android/app/src/main/AndroidManifest.xml:

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

Minimum SDK: 21

iOS #

Add the following usage descriptions to your ios/Runner/Info.plist:

<key>NSCameraUsageDescription</key>
<string>Camera access is required for face liveness detection</string>
<key>NSMicrophoneUsageDescription</key>
<string>Microphone access is required for recording video audio</string>

Minimum iOS version: 15.5

Usage #

import 'package:face_live/face_live.dart';

FaceLivenessView(
  onProgress: (progress) {
    print('Liveness progress: ${progress.toInt()}%');
  },
  onCapture: (videoFile) {
    print('Liveness success! Video: ${videoFile.path}');
    // Process the recorded video file
  },
)

How it works #

1. Camera Preview: Native camera view is embedded in Flutter using Platform Views
2. Face Detection: ML Kit processes camera frames in real-time to detect face pose
3. Progress Tracking: Measures head rotation (yaw) coverage against target span (default 60°)
4. Video Recording: Records the entire session natively for verification
5. Completion: When target head movement is achieved, returns the recorded video file

Configuration #

The FaceLivenessView widget accepts the following parameters:

• onProgress: Callback for progress updates (0.0 - 100.0)
• onCapture: Callback when liveness check succeeds with video file
• targetYawSpan: Target head rotation span in degrees (default: 60.0)
• timeoutMillis: Session timeout in milliseconds (default: 10000)

Example #

See the example/ directory for a complete implementation showing:

• Camera preview with real-time progress updates
• User guidance text
• Success handling with video file path logging

Performance #

This plugin is designed for high performance by:

• Running all heavy operations (camera, ML Kit, video encoding) natively
• Only sending lightweight progress data to Flutter
• Using hardware-accelerated camera and ML processing
• Avoiding expensive image data transfers between Dart and native layers

License #

See LICENSE file for details.