ctOS - Profiler is a mobile application inspired by the Watch Dogs series. It provides on-device real-time face recognition, tracking, and detailed target profiling using LiteRT (TensorFlow Lite).
- Features
- Tech Stack
- Quick Start
- Repository Structure
- Architecture
- Testing
- Contributing
- Common Errors
- License
- AI Face Detection & Recognition: Real-time tracking and matching using high-performance models (FaceNet, MobileFaceNet).
- Deep Profiling: Generate and edit detailed digital dossiers for targets, including:
- Demographics: Name, age, birth date (DMY format), and occupation.
- Biometrics: Height (cm) and weight (kg) tracking.
- Socio-Economic Status: Automated income level assessment.
- Risk Assessment: Real-time threat level calculation and monitoring.
- Personality Profiling: Analysis of positive and negative behavioral traits.
- Local Database: Secure on-device storage for face embeddings and profiles via sqflite.
- Framework: Flutter
- Languages: Dart, Kotlin, Java, C++
- AI Engine: LiteRT (TensorFlow Lite /
face_detection_tflite) - Computer Vision: OpenCV (
dartcv4) - Database: sqflite (SQLite)
- State Management: Riverpod (v2)
- Flutter SDK: Stable channel (>= 3.1.0)
- Android SDK: Platform 35
- Android NDK: Version 25.1.8937393 (Strictly required for native asset compilation)
- Java: Version 17
- Clone the repository:
git clone https://github.com/logm1lo/ctOS-Profiler.git- Get dependencies:
flutter pub get- Configure Environment:
Create or edit
android/local.propertiesand ensure the NDK path is explicitly set:
sdk.dir=C\:\\Users\\YourUser\\AppData\\Local\\Android\\Sdk
ndk.dir=C\:\\Users\\YourUser\\AppData\\Local\\Android\\Sdk\\ndk\\25.1.8937393
flutter.sdk=C\:\\path\\to\\flutter- Run the App:
flutter runctOS-Profiler/
├── android/ # Android platform-specific configuration
├── app/ # Main Android application module (Kotlin/Gradle)
├── assets/models/ # Pre-trained TFLite models (FaceNet, etc.)
├── buildSrc/ # Custom Gradle build logic and mock plugins
├── lib/ # Flutter source code
│ ├── core/ # Shared utilities, themes, and providers
│ ├── data/ # Data sources and repository implementations
│ ├── domain/ # Business logic, entities, and use cases
│ └── presentation/ # UI screens, widgets, and Riverpod notifiers
├── pubspec.yaml # Flutter dependencies and assets configuration
└── README.md # Project documentation
The project follows a Clean Architecture pattern with a Clear separation of concerns:
- Presentation Layer: Handles UI rendering and state management using Riverpod. Screens watch providers to react to real-time AI detections.
- Domain Layer: Defines the core entities (e.g.,
FaceEntity) and business rules (e.g.,MatchFace). - Data Layer: Implements data persistence and coordinates between local databases and native services.
- Native Infrastructure: Leverages JNI and Dart FFI to perform heavy lifting like YUV-to-RGB conversion and model inference.
If you encounter PathNotFoundException: Directory listing failed during build, ensure your ndk.dir in local.properties points to a complete NDK installation. Version 25.1.8937393 is recommended. Avoid using NDK versions that only contain an .installer folder.
This usually occurs due to a conflict in the Flutter embedding. The project uses a custom FlutterMockPlugin in buildSrc to manage these dependencies. Ensure that library sub-projects use compileOnly for the Flutter embedding.
If you see errors related to CameraControllerNotifier after dispose, ensure that any asynchronous operations (like saving profiles) are guarded with mounted checks and that the provider is not being invalidated prematurely during navigation.
This project is licensed under the GNU General Public License v3.0 - see the LICENSE file for details.