This is the updated version of the Flutter-based face authentication application. It detects faces using Google ML Kit's face detection and uses the FaceNet512 model to recognize and differentiate users. Encoded face data is stored and retrieved from Firebase Firestore.
✅ Fully compatible with:
- Flutter 3.32.6 • channel stable • flutter.git
- Framework revision 077b4a4ce1 (2025-07-08)
- Engine revision 72f2b18bb0
- Dart 3.8.1 • DevTools 2.45.1
- 👁️ Face Detection with Google ML Kit: Uses
google_mlkit_face_detectionto detect faces in real time. - 🧠 Face Encoding with FaceNet512: Converts detected faces into 512-length feature vectors using the FaceNet512 model.
- ☁️ Cloud Firestore Integration: Stores and retrieves encoded face data (Float32 arrays) and associated names.
- 🔄 Real-Time Face Authentication: Predicts user identity by comparing current face data with stored vectors using cosine similarity.
- 📦 On-device Inference with TFLite: Utilizes
tflite_flutterfor running the FaceNet model locally on the device.
cupertino_icons: ^1.0.2
camera: ^0.10.5+5
google_mlkit_face_detection: ^0.13.1
image: ^3.0.2
tflite_flutter: ^0.11.0
cloud_firestore: ^4.13.2
firebase_core: ^2.23.0
-
Clone this repository:
git clone https://github.com/tharushaudana/flutter_face_auth_2.git cd flutter_face_auth_2 -
Install dependencies:
flutter pub get
-
Set up Firebase for Android and iOS (Firebase Setup Guide).
-
Run the app:
flutter run
This project demonstrates mobile face authentication using deep learning and real-time face detection. It is suitable for secure identity verification systems where user faces are encoded, stored, and compared efficiently using on-device AI and cloud storage.