🎯 YOLOverlay

Real-time object detection overlay for macOS using YOLO models with smooth tracking, performance monitoring, and configurable settings.

✨ Features

🚀 Core Detection

• Real-time Screen Detection: Instant object detection on any screen area
• Multiple YOLO Models: Support for YOLO11 (s, l) and YOLOv8 (l, x)
• GPU Acceleration: CoreML-powered inference on Apple Silicon
• Smart Capture: Mouse cursor exclusion and frame optimization
• Configurable Detection: Adjustable confidence thresholds

🎨 Visual Experience

• Transparent Overlay: Non-intrusive detection visualization
• Customizable Display:
  • Bounding box colors and styles
  • Label font size and opacity
  • Confidence score visibility
  • Class filtering options

⚡️ Performance

• Real-time Monitoring:
  • FPS and detection latency
  • CPU/GPU utilization
  • Dropped frame detection
  • Performance logging
• Optimization Controls:
  • Configurable frame rate
  • Resource usage management
  • Detection throttling

🎛️ User Interface

• Quick Controls:
  • Status bar menu access
  • Global keyboard shortcuts
    • ⌘S: Start/Stop detection
    • ⌘L: Show performance logs
    • ⌘,: Open preferences
• Comprehensive Settings:
  • Model selection
  • Detection parameters
  • Visual customization
  • Performance tuning
  • Class management

🔧 Developer Tools

• Advanced Logging:
  • Real-time log streaming
  • Log level filtering
  • Millisecond precision
  • Subsystem isolation
• Performance Metrics:
  • Detection statistics
  • Resource monitoring
  • Debug information

🖥️ Requirements

• macOS 15.1 or later
• Apple Silicon Mac (M1 or newer)
• Screen Recording permission

🚀 Installation

1. Download latest release
2. Move to Applications
3. Grant Screen Recording permission
4. Launch and customize

📖 Usage

1. Click menu bar eye icon or ⌘S to start
2. Adjust settings with ⌘,
3. Monitor performance with ⌘L
4. Filter and customize as needed

🤖 Adding Custom Models

YOLOverlay supports YOLO models in CoreML format (.mlpackage). To add a new model:

1. Convert YOLO Model:

   # Create Python virtual environment
   python -m venv .venv
   source .venv/bin/activate
   
   # Install requirements
   pip install -r requirements.txt
   
   # Convert model (e.g., YOLOv8n)
   ./convert_yolo.py v8 n

   This will create a .mlpackage file in the project directory.

2. Add to Xcode Project:

   • Open the project in Xcode
   • Right-click on the YOLOverlay group
   • Select "Add Files to YOLOverlay..."
   • Choose the .mlpackage file
   • Ensure "Copy items if needed" is checked
   • Click "Add"

3. Build and Run:

   • Clean the build folder (Shift+⌘+K)
   • Build the project (⌘B)
   • Run the app (⌘R)

The new model will appear in the Model tab of Preferences.

🏆 Credits

This project is built on powerful technologies:

Object Detection

• YOLO11 by Ultralytics (Glenn Jocher & Jing Qiu)
  • State-of-the-art object detection
  • Version 11.0.0 (2024)
  • View on GitHub

Apple Technologies

• Vision Framework: Advanced computer vision and image analysis
• Core ML: On-device machine learning inference
• ScreenCaptureKit: High-performance screen capture

---

Made with ❤️ by kazazes