Face Recognition Camera(FRC) #

Setup #

• Download model here

• Add model mobile_face_net.flite in pubspec.yaml

  assets:
    - assets/

• Create a function to save and read in locale storage. In this case using shared_preferences packages.

flutter pub add shared_preferences

or

dependencies:  
  # add shared_preferences to your dependencies
  shared_preferences:

Example code:

import 'dart:convert';
import 'dart:typed_data';

import 'package:shared_preferences/shared_preferences.dart';

class LocaleStorage  {
  static LocaleStorage? _instance;
  static late SharedPreferences _sharedPref;

  LocaleStorage() {
    init();
  }

  static Future<LocaleStorage> init() async {
    if (_instance != null) return _instance!;
    _sharedPref = await SharedPreferences.getInstance();
    return _instance = LocaleStorage();
  }

  Uint8List? read(String key) {
    try {
      final data = _sharedPref.getString(key);
      if (data == null) return null;
      return base64Decode(data);
    } catch (e) {
      throw Exception('Failed to read ${e.toString()}');
    }
  }

  Future<void> write(String key, List<double> embedding) async {
    try {
      String base64String = base64Encode(byteData(embedding));
      await _sharedPref.setString(key, base64String);
    } catch (e) {
      throw Exception('Failed to write ${e.toString()}');
    }
  }

  Uint8List byteData(List<double> embedding) {
    final value = Uint8List.fromList(
      embedding.map((e) => (e * 255).clamp(0, 255).toInt()).toList(),
    );
    return value;
  }
}

Usage #

• Initialize in main.dart

void main() async {
  WidgetsFlutterBinding.ensureInitialized();
  await Frc.initialize();
  await LocaleStorage().init();
  runApp(const MyApp());
}

• Load model from asset

RecognizerHandler('assets/mobile_face_net.tflite')

• Save image to local as Uint8List

LocaleStorage().write('sample', uint8ListToListDouble(img));

List<double> uint8ListToListDouble(Uint8List? uint8List) {
    if (uint8List == null || uint8List.isEmpty) return [];
    return uint8List.map((e) => e / 255.0).toList();
  }

• Compare image from input and locale

final localeImage = LocaleStorage().read('sample');

await RecognizerHandler()compareImages(input, localeImage);

• Initialize Controller

FRCController(
  onCapture: (image, img) {
    setState(() {
      _capturedImage = image;
      imageFile = img;
    });
  },
);

• use FaceCamera

FaceCamera(
  controller: controller,
  showCaptureControl: true,
  captureControl: (file, img) {}
)

• Use FaceCameraCircle

FaceCameraCircle(
  controller: controller,
  autoDisableCaptureControl: true,
) 

• Don't forget to set value autoCapture = true and centerPosition: true when using FaceCameraCircle

See full example here.