Face Liveness Detection Plus #

A real-time facial verification package for Flutter using Google ML Kit for liveness detection. It ensures user interaction through smiling, blinking, and head movements. Key capabilities include real-time face detection, dynamic UI feedback, countdown timers, manual/programmatic capture control via FaceCaptureController, and image path retrieval for secure authentication and anti-spoofing verification.

Features #

• Real-Time Detection: Fast and accurate face detection powered by Google ML Kit with configurable performanceMode (fast or accurate).
• Configurable Thresholds: Customize sensitivity thresholds for smiling, blinking, and head movements via LivenessThresholds.
• Localization & Instructions: Easily customize or translate instruction texts via LivenessLocalization.
• Custom UI Overlay: Replace default dotted border with custom customOverlayBuilder.
• Liveness Controller: Isolated logic and state controller LivenessDetectionController for testing and custom UI integration.
• Dynamic Feedback: Real-time visual UI feedback for each verification rule.
• Animated Transitions: Smooth progress animations during rule evaluation.
• Countdown Timer: Built-in countdown timer before verification completes.
• Capture Controller: Flexible manual or automatic image capture control (FaceCaptureController).
• Accuracy Calculation: Computes confidence/accuracy percentage per completed rule.

Installation #

Add face_liveness_detection_plus to your pubspec.yaml:

dependencies:
  face_liveness_detection_plus: ^1.2.0

Run flutter pub get to install the package.

Platform Setup #

iOS #

Set your global platform target in ios/Podfile:

platform :ios, '15.5'

Add camera and microphone permission descriptions to ios/Runner/Info.plist:

<key>NSCameraUsageDescription</key>
<string>Camera access is required for real-time face liveness detection.</string>
<key>NSMicrophoneUsageDescription</key>
<string>Microphone access is required for video recording during verification.</string>

Preview #

Usage Example #

Basic Verification View #

import 'package:flutter/material.dart';
import 'package:flutter/cupertino.dart';
import 'package:face_liveness_detection_plus/face_liveness_detection_plus.dart';

class FaceVerificationWidget extends StatefulWidget {
  const FaceVerificationWidget({super.key});

  @override
  State<FaceVerificationWidget> createState() => _FaceVerificationWidgetState();
}

class _FaceVerificationWidgetState extends State<FaceVerificationWidget> {
  final List<Rulesets> _completedRuleset = [];

  @override
  Widget build(BuildContext context) {
    return Scaffold(
      body: FaceDetectorView(
        onSuccessValidation: (validated) {},
        onValidationDone: (controller) => const Center(
          child: Text('Verification Complete!'),
        ),
        onRulesetCompleted: (ruleset, imageUrl) {
          if (!_completedRuleset.contains(ruleset)) {
            setState(() => _completedRuleset.add(ruleset));
          }
        },
        child: ({required countdown, required state, required hasFace}) {
          return Column(
            children: [
              const SizedBox(height: 20),
              Row(
                mainAxisAlignment: MainAxisAlignment.center,
                children: [
                  const Icon(Icons.face, size: 30),
                  const SizedBox(width: 10),
                  Text(
                    hasFace ? 'User face found' : 'User face not found',
                    style: _textStyle,
                  ),
                ],
              ),
              const SizedBox(height: 30),
              Text(
                _rulesetHints[state] ?? 'Please follow instructions',
                style: _textStyle.copyWith(fontSize: 20, fontWeight: FontWeight.w600),
              ),
              if (countdown > 0)
                Text(
                  'Timer: $countdown',
                  style: _textStyle.copyWith(fontSize: 16),
                )
              else
                const CupertinoActivityIndicator(),
            ],
          );
        },
      ),
    );
  }
}

const TextStyle _textStyle = TextStyle(
  color: Colors.black,
  fontWeight: FontWeight.w400,
  fontSize: 12,
);

const Map<Rulesets, String> _rulesetHints = {
  Rulesets.smiling: 'Please Smile',
  Rulesets.blink: 'Please Blink',
  Rulesets.tiltUp: 'Please Look Up',
  Rulesets.tiltDown: 'Please Look Down',
  Rulesets.toLeft: 'Please Look Left',
  Rulesets.toRight: 'Please Look Right',
};

Manual Capture using FaceCaptureController #

final FaceCaptureController _controller = FaceCaptureController();

FaceDetectorView(
  autoCapture: false,
  controller: _controller,
  onRulesetCompleted: (rule, imageUrl) {
    print('Completed rule: $rule, image: $imageUrl');
  },
  onValidationDone: (controller) => Container(),
  child: ({required countdown, required state, required hasFace}) {
    return ElevatedButton(
      onPressed: () async {
        final result = await _controller.capture(null);
        print('Captured ${result.rule} with accuracy ${result.accuracyPercentage}%');
      },
      child: const Text('Capture Image'),
    );
  },
)

Credits #

This package is based on and inspired by the original work of Roshan Karki (facelivenessdetection). Special thanks for the core implementation of Flutter face liveness detection.

License #

This project is licensed under the MIT License.