TagMyStockUpscaler

A native macOS app that uses Real-ESRGAN to upscale images with AI-powered enhancement. Built specifically for ARM-based Macs (Apple Silicon) with SwiftUI.

Features

• 🖼️ Drag & Drop Interface: Simply drag images into the app or use the file picker
• 🚀 4x Upscaling: Uses Real-ESRGAN_x4plus model for high-quality image enhancement
• 🎯 ARM Native: Optimized for Apple Silicon Macs (M1, M2, M3, etc.)
• 💾 Export Options: Save upscaled images in PNG or JPEG format
• 📱 Modern UI: Clean SwiftUI interface with progress tracking
• 🔧 Auto Setup: Automatically downloads Real-ESRGAN if not installed

What is Real-ESRGAN?

Real-ESRGAN is a state-of-the-art AI model developed by researchers at Tencent ARC Lab that specializes in:

• Upscaling low-resolution images to higher resolutions
• Enhancing image quality and reducing artifacts
• Working exceptionally well with photos, artwork, and illustrations
• Maintaining sharp details while increasing image size

Research Paper: Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data

Requirements

• macOS 12.0 or later
• Apple Silicon Mac (M1/M2/M3) - recommended
• Intel Mac - supported but may have different performance
• Xcode 14+ (for building from source)

Installation

Option 1: Build from Source

1. Clone this repository:

git clone https://github.com/yourusername/TagMyStockUpscaler.git
cd TagMyStockUpscaler

2. Open TagMyStockUpscaler.xcodeproj in Xcode

3. Build and run the project (⌘+R)

4. The app will automatically guide you through setting up Real-ESRGAN on first launch

Option 2: Manual Real-ESRGAN Setup

If you prefer to set up Real-ESRGAN manually before running the app:

# Run the included setup script
./setup_realesrgan.sh

Or install via Homebrew:

brew install xinntao/real-esrgan/realesrgan-ncnn-vulkan

Usage

1. Launch the App: Open TagMyStockUpscaler from Applications or Xcode
2. Setup (First Time): If Real-ESRGAN isn't installed, follow the setup guide
3. Upload Image: Drag & drop an image or click "Select Image"
4. Upscale: Click "Upscale Image" and wait for processing
5. Save: Click "Save Upscaled Image" to export the result

Supported Image Formats

Input: JPEG, PNG, TIFF, HEIC, WebP, and most common image formats Output: PNG, JPEG

Tips for Best Results

• Photos: Works great with portraits, landscapes, and general photography
• Artwork: Excellent for digital art, illustrations, and drawings
• Screenshots: Perfect for enhancing low-resolution screenshots
• Avoid: Very high-resolution images (may be slow), heavily compressed images

Technical Details

• Model: Real-ESRGAN_x4plus (optimized for general images)
• Upscaling Factor: 4x (can output at custom scales)
• Backend: NCNN with Vulkan acceleration
• Processing: Local processing (no internet required after setup)
• Memory: Processes images in tiles to handle large images efficiently

Architecture

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   SwiftUI App   │ -> │  ImageUpscaler   │ -> │ Real-ESRGAN     │
│                 │    │     Class        │    │   Executable    │
└─────────────────┘    └──────────────────┘    └─────────────────┘

• ContentView.swift: Main UI with drag/drop and image display
• ImageUpscaler.swift: Handles Real-ESRGAN integration and processing
• SetupView.swift: Guides users through Real-ESRGAN installation

Troubleshooting

Real-ESRGAN Not Found

If you see "Real-ESRGAN not found":

1. Use the automatic download in the app
2. Run ./setup_realesrgan.sh manually
3. Install via Homebrew (see installation section)

Permission Errors

macOS may ask for permission to run Real-ESRGAN:

1. Go to System Preferences → Security & Privacy
2. Allow the Real-ESRGAN executable to run
3. Or right-click the executable and select "Open"

Slow Processing

If processing is slow:

• Ensure you're on an Apple Silicon Mac for best performance
• Try with smaller images first
• Close other intensive applications

Memory Issues

For very large images:

• The app uses tiling to process large images
• If you encounter issues, try with smaller source images
• Ensure you have sufficient free disk space

Performance

Processing times on M1 MacBook Pro (8GB RAM):

• 512x512 image: ~5-10 seconds
• 1024x1024 image: ~15-30 seconds
• 2048x2048 image: ~60-120 seconds

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly on Apple Silicon
5. Submit a pull request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Real-ESRGAN is licensed under the BSD 3-Clause License by Tencent ARC Lab.

Acknowledgments

• Real-ESRGAN: xinntao/Real-ESRGAN - The amazing AI upscaling technology
• NCNN: Tencent/ncnn - High-performance neural network inference framework
• Apple: For the excellent SwiftUI framework and Apple Silicon performance

Support

If you encounter any issues:

1. Check the troubleshooting section above
2. Look at existing GitHub Issues
3. Create a new issue with:
   • macOS version
   • Mac model (Intel/Apple Silicon)
   • Steps to reproduce
   • Error messages or screenshots

---

Made with ❤️ for the macOS community