face_verify 0.0.3

face_verify: ^0.0.3 copied to clipboard

A Flutter package for face detection and recognition using Google ML Kit and TensorFlow Lite.

Features #

• Face Detection: Detect faces in images using Google ML Kit.
• Face Recognition: Recognize faces by comparing embeddings with registered users.
• Customizable Overlays: Display custom overlays around detected faces.
• Performance Optimization: Customizable resolution of the camera feed and performance mode of the face detector.
• Offline Capability: Perform face detection and recognition without an internet connection.

Getting Started #

Installation #

1. Add the package to your project:

flutter pub add face_verify

2. Import the package in your Dart code:

import 'package:face_verify/face_verify.dart';

Usage #

Initialize Cameras #

Future<void> main() async {
  WidgetsFlutterBinding.ensureInitialized();
  final cameras = await availableCameras();
  // choose which camera you want to use
  final cameraDescription = cameras.first;
}

Register Users #

  final List<UserModel> users = await registerUsers(
    registerUserInputs: [
      RegisterUserInputModel(name: 'User1', imagePath: 'path/to/image1.jpg'),
      RegisterUserInputModel(name: 'User2', imagePath: 'path/to/image2.jpg'),
    ],
    cameraDescription: cameraDescription,
  );

Note

The images provided to register users must be a close-up photo with a white background, similar to a passport or ID image, and should be of high quality. This can significantly affect the model's efficiency and accuracy.

Perform Face Recognition #

  Set<UserModel>? recognizedUsers = await Navigator.push(
                      context,
                      MaterialPageRoute(
                        builder: (context) => DetectionView(
                          users: users,
                          cameraDescription: cameraDescription,
                        ),
                      ));
                  if (recognizedUsers != null && recognizedUsers.isNotEmpty) {
                    log('recognized users : $recognizedUsers');
                  }

Parameters #

The DetectionView widget provides several parameters that can be customized to adjust the behavior and appearance of the face detection and recognition process:

Authors #

• Badieh Nader - GitHub

Acknowledgments #

This package utilizes a FaceNet model for face recognition, as described in the paper "FaceNet: A Unified Embedding for Face Recognition and Clustering" (Schroff et al., 2015). The model has been converted to TensorFlow Lite format. Model source: Unknown.