Face Blur Pro

🤖 Built with Antigravity + Gemini 3 Model — This app was created using Vibe Coding, an AI-assisted development approach powered by Gemini 3.

☁️ Google Cloud credits are provided for this project. #AISprintH2

Face Blur Pro is a cross-platform Flutter application designed to enhance privacy by detecting and blurring faces in images. It demonstrates a powerful hybrid architecture, leveraging on-device ML for mobile, ONNX Runtime Web for browser, and a Rust-based backend for desktop performance.

✨ Key Features

• Cross-Platform Support: Runs smoothly on iOS, Android, macOS, Windows, Linux, and Web.
• Hybrid Face Detection:
  • Mobile (Android/iOS): Utilizes Google ML Kit for fast, efficient on-device detection.
  • Desktop (macOS/Windows/Linux): Powered by a Rust backend using tract-onnx and the version-RFB-640.onnx model.
  • Web: Uses ONNX Runtime Web with the same version-RFB-640.onnx model for consistent detection across platforms.
• Manual Region Selection: Draw custom blur regions when face detection misses faces. Supports drag-to-draw and long-press to delete.
• Privacy-First: All processing is performed locally on your device. No images are ever uploaded to the cloud.
• Rust-Powered Image Processing: Uses the image crate in Rust for high-performance pixel manipulation and blurring.
• Flexible Blurring: Choose between Rectangular or Circular blur shapes.
• Share & Save: Easily save your edited images to the gallery or share them directly with other apps.

🖥️ Supported Platforms

Platform	Face Detection	Blur Processing	Status
iOS	Google ML Kit	Rust FFI	✅ Supported
Android	Google ML Kit	Rust FFI	✅ Supported
macOS	ONNX (Rust)	Rust FFI	✅ Supported
Windows	ONNX (Rust)	Rust FFI	✅ Supported
Linux	ONNX (Rust)	Rust FFI	✅ Supported
Web	ONNX Runtime Web	Canvas API	✅ Supported

🛠️ Tech Stack

• Frontend: Flutter (Dart)
• Backend: Rust
• Bridge: flutter_rust_bridge connects Dart and Rust.
• Mobile ML: google_mlkit_face_detection
• Desktop ML: tract-onnx (ONNX Runtime in pure Rust)
• Web ML: ONNX Runtime Web with version-RFB-640.onnx (RetinaFace variant)
• Image Processing: image crate (Rust) / Canvas API (Web)

🚀 Installation & Setup

Prerequisites

Ensure you have the following installed on your system:

1. Flutter SDK: Install Flutter
2. Rust Toolchain: Install Rust
3. C++ Build Tools:
   • macOS: Xcode (install via App Store or xcode-select --install)
   • Windows: Visual Studio 2022 with "Desktop development with C++" workload.
   • Linux: build-essential, pkg-config, libgtk-3-dev (depending on your distro).

Steps

1. Clone the Repository:

   git clone https://github.com/jaichangpark/face_blur_pro.git
   cd face_blur_pro

2. Install Flutter Dependencies:

   flutter pub get

3. Run the App:

   # iOS/Android
   flutter run
   
   # macOS
   flutter run -d macos
   
   # Web
   flutter run -d chrome

📱 Usage Guide

1. Pick an Image: Tap the 📷 사진 열기 button to select a photo from your device.
2. Detect Faces: The app will automatically run face detection. Detected faces will be highlighted with red boxes.
3. Select Faces: Tap on any box to toggle selection. Selected faces will turn green.
4. Manual Drawing (New!):
   • Tap the ✏️ Edit icon to enter drawing mode.
   • Drag on the image to draw custom blur regions (shown in blue).
   • Long-press to delete manually added regions.
5. Choose Blur Style: Toggle between Circle and Rectangle icons to change the blur shape.
6. Apply Blur: Tap the 블러 실행 button. The selected areas will be blurred.
7. Save/Share:
   • Tap the 💾 Save icon to save the image to your gallery.
   • Tap the 🔗 Share icon to send the image to another app.

📂 Project Structure

face_blur_pro/
├── lib/                          # Flutter (Dart) code
│   ├── main.dart                 # Entry point
│   ├── data/models/              # Data models (MyFace, BlurShape)
│   ├── domain/services/          # Business logic services
│   │   ├── image_processing_service.dart
│   │   ├── web_face_service.dart       # Web platform conditional import
│   │   ├── web_face_service_web.dart   # Web implementation
│   │   └── web_face_service_stub.dart  # Native stub
│   └── presentation/home/        # UI components
├── rust/                         # Rust backend
│   └── src/api/simple.rs         # Face detection & blur logic
├── web/                          # Web-specific files
│   ├── index.html                # ONNX Runtime Web scripts
│   └── face_blur_web.js          # JS face detection logic
└── assets/models/                # ONNX models
    └── version-RFB-640.onnx      # RetinaFace variant model

📄 License

This project is open-source. See the LICENSE file for details.