PixelEraser Pro

Professional background removal and image editing tool with AI-powered upscaling

Intelligent background removal • Precise editing tools • AI upscaling

Professional-grade image editing with a DaVinci Resolve-inspired dark interface

Download Latest Release | Report Bug | Request Feature

---

Table of Contents

• For Users
• For Developers
• Additional Info

---

For Users

Table of Contents

1. Features Overview
2. Getting Started
3. Tools & Features
   • Auto Color Remove
   • Manual Eraser
   • Repair Tool
   • AI Upscaling
4. Navigation & Controls
5. Export Options
6. Keyboard Shortcuts
7. Tips & Best Practices

---

1. Features Overview

PixelEraser Pro is a desktop application designed for professional background removal and image editing. Built with Qt and OpenCV, it offers powerful tools with an intuitive interface.

Key Features:

Feature	Description
Auto Color Remove	Intelligent background removal using LAB color space analysis
Manual Eraser	Precise control with adjustable brush size and edge softening
Repair Tool	Fix over-erased areas with smart restoration
AI Upscaling	12 upscaling models including Real-ESRGAN and Real-CUGAN
Undo/Redo	Up to 10 levels of history for safe editing
Zoom & Pan	Smooth navigation with Space bar panning
Compare Mode	Press H to compare with original image
Export Options	PNG, JPEG, WebP with quality control

---

2. Getting Started

System Requirements:

Requirement	Specification
Operating System	Windows 10/11 64-bit
RAM	4GB minimum, 8GB recommended
Graphics	OpenGL support
Disk Space	500MB for installation

Installation:

1. Download PixelEraserPro_Setup_v1.0.0.exe from Releases
2. Run the installer
3. Follow the installation wizard
4. Launch PixelEraser Pro from Start Menu or Desktop

First Launch:

1. Click File → Open (Ctrl+O) or drag and drop an image
2. Supported formats: PNG, JPEG, JPG, BMP, WebP
3. The image will load in the canvas
4. Start editing with the tools in the sidebar

---

3. Tools & Features

Auto Color Remove

Intelligent background removal that analyzes colors in the LAB color space for accurate selection.

How to use:

1. Click the Auto Color Remove button or press A
2. Click on the background color you want to remove
3. Adjust tolerance (0-100) to control selection sensitivity
4. The tool removes similar colors in the visible viewport

Settings:

Setting	Range	Description
Tolerance	0-100	Higher values remove more similar colors
Edge Softening	0-10	Smooths edges for natural results

Tips:

• Start with tolerance around 50
• For complex backgrounds, use multiple clicks
• Zoom in for precise selection
• Only processes visible area when zoomed

---

Manual Eraser

Precise eraser tool for manual background removal with adjustable brush size.

How to use:

1. Click the Eraser button or press E
2. Adjust brush size (1-200 pixels)
3. Click and drag to erase
4. Use edge softening for smooth transitions

Settings:

Setting	Range	Description
Brush Size	1-200	Diameter of the eraser brush
Edge Softening	0-10	Feathers the brush edges

Tips:

• Use smaller brushes for detailed work
• Increase edge softening for natural results
• Hold Space to pan while erasing
• Use Ctrl+Z to undo mistakes

---

Repair Tool

Restore over-erased areas by bringing back the original image.

How to use:

1. Click the Repair button or press R
2. Adjust brush size
3. Paint over areas you want to restore
4. The original image will be revealed

Use cases:

• Fix accidentally erased areas
• Restore details after aggressive auto removal
• Blend edges between erased and kept areas

---

AI Upscaling

Enhance image resolution using state-of-the-art AI models. Choose from 12 different models optimized for various image types.

Available Models:

Category	Models	Best For
General Purpose	Real-ESRGAN x4/x2 Plus	Photos and realistic images
Anime & Illustrations	Real-ESRGAN x4 Anime	Anime, manga, and drawn art
Denoise	Real-ESRGAN x4/x2 Denoise	Removing JPEG artifacts and noise
Conservative	Real-CUGAN 2x/3x/4x Conservative	Minimal changes, fewer artifacts
No Denoise	Real-CUGAN 2x/3x/4x No Denoise	Preserving film grain and texture

How to use:

1. Open an unmodified image (upscaling only works on original images)
2. Go to Image → Upscale (Ctrl+Alt+U)
3. Select a model from the dropdown
4. Click Download Model if not already downloaded (~17MB)
5. Click Upscale and wait for processing
6. The upscaled image will replace the current image

Model Selection Guide:

Image Type	Recommended Model
Photos	Real-ESRGAN x4 Plus
Compressed photos	Real-ESRGAN x4 Denoise
Anime/Manga	Real-ESRGAN x4 Anime
Film scans	Real-CUGAN 4x No Denoise
General upscaling	Real-CUGAN 4x Conservative

Notes:

• Models are downloaded once and cached locally
• Processing time depends on image size
• Larger images are processed in tiles
• Upscaling only works on images without transparency

---

4. Navigation & Controls

Zoom:

Action	Method
Zoom In	Mouse wheel up / Ctrl + Plus
Zoom Out	Mouse wheel down / Ctrl + Minus
Fit to Screen	Double-click canvas / Ctrl+0
Actual Size	Ctrl+1

Pan:

Action	Method
Pan	Hold Space + Drag
Pan (alternative)	Middle mouse button + Drag

View:

Action	Shortcut
Toggle Sidebar	Tab
Compare with Original	Hold H
Checkerboard Background	Automatic for transparency

---

5. Export Options

Export your edited images in various formats with quality control.

How to export:

1. Click File → Export (Ctrl+E)
2. Choose format: PNG, JPEG, or WebP
3. Adjust quality settings
4. Select save location
5. Click Export

Format Comparison:

Format	Transparency	Quality	File Size	Best For
PNG	Yes	Lossless	Large	Final output, transparency needed
JPEG	No	Lossy	Small	Photos, no transparency
WebP	Yes	Lossy/Lossless	Medium	Web use, modern browsers

Quality Settings:

Format	Range	Recommendation
PNG	N/A	Always lossless
JPEG	1-100	85-95 for high quality
WebP	1-100	80-90 for balanced quality/size

---

6. Keyboard Shortcuts

File Operations:

Shortcut	Action
Ctrl+O	Open Image
Ctrl+E	Export
Ctrl+Q	Quit

Editing:

Shortcut	Action
Ctrl+Z	Undo
Ctrl+Y	Redo
A	Auto Color Remove
E	Eraser Tool
R	Repair Tool

View:

Shortcut	Action
Space	Pan (hold and drag)
H	Compare with Original (hold)
Tab	Toggle Sidebar
Ctrl+0	Fit to Screen
Ctrl+1	Actual Size
Ctrl + Plus	Zoom In
Ctrl + Minus	Zoom Out

Image:

Shortcut	Action
Ctrl+Alt+R	Resize Image
Ctrl+Alt+U	Upscale Image

Help:

Shortcut	Action
F1	Keyboard Shortcuts

---

7. Tips & Best Practices

For Best Results:

Tip	Why
Start with high-resolution images	Better quality after editing
Use Auto Color Remove first	Faster than manual erasing
Adjust tolerance gradually	Prevents over-removal
Enable edge softening	Creates natural-looking edges
Zoom in for details	More precise editing
Use Compare mode (H)	Check your progress
Save frequently	Use Export to save your work

Performance Tips:

Tip	Benefit
Close other applications	More RAM for processing
Work on smaller images first	Faster processing
Use Fit to Screen	Better overview
Limit undo history	Saves memory

Common Mistakes to Avoid:

Mistake	Solution
Too high tolerance	Start low, increase gradually
Forgetting to zoom	Zoom in for precise work
Not using edge softening	Enable for natural edges
Upscaling edited images	Upscale original images only

---

For Developers

This section is for developers who want to build from source or contribute to the project.

Development Setup

Prerequisites:

Requirement	Version
Qt	6.10.1
CMake	3.20+
C++ Compiler	MSVC 2022 (Windows)
vcpkg	Latest
OpenCV	4.x
ONNX Runtime	1.x

Installation:

1. Clone the repository:

git clone https://github.com/RohitPoul/PixelEraserPro.git
cd PixelEraserPro

2. Install dependencies via vcpkg:

vcpkg install opencv4:x64-windows
vcpkg install onnxruntime:x64-windows

3. Open in Qt Creator:
   • File → Open File or Project
   • Select CMakeLists.txt
   • Configure with MSVC 2022 64-bit kit
   • Build (Ctrl+B)

---

Project Structure

PixelEraserPro/
├── include/                    # Header files
│   ├── MainWindow.h
│   ├── CanvasWidget.h
│   ├── ImageProcessor.h
│   ├── ToolManager.h
│   ├── HistoryManager.h
│   ├── Upscaler.h
│   ├── UpscaleDialog.h
│   ├── ExportDialog.h
│   ├── ResizeDialog.h
│   └── UpdateChecker.h
├── src/                        # Source files
│   ├── main.cpp
│   ├── MainWindow.cpp
│   ├── CanvasWidget.cpp
│   ├── ImageProcessor.cpp
│   ├── ToolManager.cpp
│   ├── HistoryManager.cpp
│   ├── Upscaler.cpp
│   ├── UpscaleDialog.cpp
│   ├── ExportDialog.cpp
│   ├── ResizeDialog.cpp
│   └── UpdateChecker.cpp
├── resources/                  # Resources
│   ├── icons/
│   │   ├── app-icon.ico
│   │   └── app-icon.png
│   ├── app.rc
│   └── resources.qrc
├── libs/                       # Third-party libraries
│   ├── stb_image.h
│   └── stb_image_write.h
├── build/                      # Build output
├── CMakeLists.txt             # CMake configuration
├── installer.iss              # Inno Setup script
├── deploy.ps1                 # Deployment script
└── README.md

---

Building the Application

Debug Build:

cmake -S . -B build -DCMAKE_BUILD_TYPE=Debug
cmake --build build --config Debug

Release Build:

cmake -S . -B build -DCMAKE_BUILD_TYPE=Release
cmake --build build --config Release

Using Qt Creator:

1. Open project in Qt Creator
2. Select Release mode (bottom-left)
3. Build → Rebuild All (Ctrl+B)

---

Creating a Release

Automated Deployment:

1. Build in Release mode
2. Run the deployment script:

cd PixelEraserPro
.\deploy.ps1

This will:

• Create portable version in ../PixelEraserPro_Portable
• Copy all required DLLs (Qt, OpenCV, ONNX Runtime)
• Create installer using Inno Setup
• Output: ../PixelEraserPro_Setup_v1.0.0.exe

Manual Deployment:

1. Build in Release mode
2. Run windeployqt on the executable:

windeployqt.exe PixelEraserPro.exe

3. Copy OpenCV DLLs from vcpkg
4. Copy ONNX Runtime DLL from vcpkg
5. Create installer with Inno Setup

---

Architecture

Design Pattern: Model-View-Controller (MVC)

Component	Responsibility
MainWindow	Main UI, menu bar, toolbar, status bar
CanvasWidget	Image display, zoom, pan, tool interaction
ImageProcessor	Image manipulation, color removal, erasing
ToolManager	Tool state, brush settings, tool switching
HistoryManager	Undo/redo stack, state management
Upscaler	AI upscaling, ONNX Runtime integration
UpdateChecker	GitHub release checking, auto-updates

Key Technologies:

Technology	Purpose
Qt 6	GUI framework, cross-platform support
OpenCV	Image processing, color space conversion
ONNX Runtime	AI model inference for upscaling
CMake	Build system
vcpkg	Dependency management

---

Contributing

Contributions are welcome! Here's how you can help:

Bug Reports:

1. Check if the issue already exists
2. Create a new issue with:
   • Clear description
   • Steps to reproduce
   • Expected vs actual behavior
   • Screenshots if applicable
   • System information

Feature Requests:

1. Describe the feature
2. Explain the use case
3. Provide examples if possible

Pull Requests:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request with clear description

Code Style:

• Follow Qt coding conventions
• Use meaningful variable names
• Comment complex logic
• Keep functions focused and small

---

Additional Info

Technical Specifications

Image Processing:

Feature	Implementation
Color Space	LAB for color removal, RGB for display
Flood Fill	Custom implementation with tolerance
Edge Softening	Gaussian blur on alpha channel
Large Image Optimization	Lazy rendering for images >8.3MP

AI Upscaling:

Feature	Details
Framework	ONNX Runtime
Models	Real-ESRGAN, Real-CUGAN
Processing	Tile-based (512x512) with padding
Formats	ONNX format (.onnx files)
Cache Location	%LOCALAPPDATA%\PixelEraserPro\models

Performance:

Metric	Value
Max Undo Levels	10
Large Image Threshold	8.3 megapixels (4K)
Tile Size (Upscaling)	512x512 pixels
Tile Padding	10 pixels

---

Supported Formats

Input Formats:

Format	Extension	Notes
PNG	.png	Full support, transparency
JPEG	.jpg, .jpeg	No transparency
BMP	.bmp	Windows bitmap
WebP	.webp	Modern format

Export Formats:

Format	Extension	Transparency	Compression
PNG	.png	Yes	Lossless
JPEG	.jpg	No	Lossy
WebP	.webp	Yes	Lossy/Lossless

---

License

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

---

Acknowledgments

Third-Party Libraries:

Library	Purpose	License
Qt	GUI framework	LGPL v3
OpenCV	Image processing	Apache 2.0
ONNX Runtime	AI inference	MIT
stb_image	Image loading	Public Domain
Real-ESRGAN	Upscaling models	BSD-3-Clause

Special Thanks:

• xinntao for Real-ESRGAN models
• bilibili for Real-CUGAN models
• Qt Company for the excellent framework
• OpenCV community for image processing tools

---

Contact & Support

Developer: Rohit Poul

Connect:

Platform	Link
GitHub	@RohitPoul
Email	poulrohit258@gmail.com

Support:

Platform	Link
GitHub Issues	Report Bug
GitHub Discussions	Ask Questions

Other Projects:

• Telegram Sticker Maker

---

⭐ Star this repo if you find it useful!