IFG-Based Contrast Enhancement

A standalone, cross-platform application for low-contrast color image enhancement using the Intuitionistic Fuzzy Generator (IFG) approach.

Overview

This application enhances low-light and low-contrast images using a two-stage pipeline:

1. Intuitionistic Fuzzy Generator (IFG)
   Transforms the luminance component of an image using fuzzy membership, non-membership, and hesitation principles, while optimizing the parameter k using entropy maximization.

2. Contrast Limited Adaptive Histogram Equalization (CLAHE)
   Improves local contrast while suppressing noise.

The GUI provides:

• Interactive image loading (including drag-and-drop)
• CLAHE and IFG enhancement
• Side-by-side comparison (Original vs IFG, or CLAHE vs IFG)
• Zooming, panning, and divider-based comparison
• Exporting enhanced output images

Installation

From Source

Clone this repository and install dependencies:

git clone [email protected]:Adi8712/IFGContrastEnhancer.git
cd IFGContrastEnhancer
python -m venv .venv
source .venv/bin/activate       # or .venv\Scripts\activate on Windows
pip install -r requirements.txt

Run the application:

python main.py

From Executable (PyInstaller Build)

If you built a standalone executable:

./dist/IFGContrastEnhancer        # Linux/macOS
dist/IFGContrastEnhancer.exe      # Windows

Usage

1. Launch the application

   python main.py

   or run the packaged executable.

2. Load an image

   • Opens inside the project's samples/ directory by default
   • Supported formats: .png, .jpg, .jpeg, .bmp, .tiff
   • You may also drag-and-drop files onto the window

3. Run Enhancement

   • Click Run Enhancement
   • CLAHE and IFG-enhanced images will be generated
   • The optimized k value is shown in the status bar

4. Compare results

   • Use the "Compare" tab for side-by-side comparison
   • Switch between "Original vs IFG" and "CLAHE vs IFG"
   • Drag the divider for interactive region comparison

5. Save output

   • Export the IFG-enhanced result as PNG or JPEG

Project Structure

.
├── build.sh                      # Cross-platform PyInstaller build helper
├── main.py                       # Application entrypoint
├── pyproject.toml                # Project metadata and dependencies
├── requirements.txt              # Pinned package versions
├── samples/                      # Optional sample input images
└── src/
    ├── enhancements/
    │   ├── clahe.py              # CLAHE enhancement implementation
    │   └── ifg.py                # IFG enhancement algorithm
    ├── gui/
    │   ├── image_views.py        # Image display and comparison widgets
    │   ├── worker.py             # Background processing thread
    │   └── main_window.py        # Main GUI layout and actions
    └── utils/
        └── resource.py           # PyInstaller-safe resource path handling

Dependencies

The application relies on:

• Python
• PySide6
• OpenCV (opencv-python)
• NumPy

Install all dependencies with:

pip install -r requirements.txt

Building a Standalone Executable

The project includes a platform-independent build.sh script that works on:

• Linux
• macOS
• Windows (via Git Bash, MSYS2, or WSL)

It automatically:

• Cleans old build artifacts
• Packages the app using PyInstaller
• Bundles the samples/ directory if present

Run the build

chmod +x build.sh
./build.sh

The final executable will appear in:

dist/IFGContrastEnhancer      # Linux/macOS
dist/IFGContrastEnhancer.exe  # Windows

Samples

The samples/ directory contains optional demo images. When using the packaged executable, these images are bundled automatically (if present during build).

Google Drive – LOw Light paired dataset (LOL)

These images are provided for educational and evaluation purposes.

Credits and Acknowledgments

This application is an open-source implementation of the image enhancement algorithm described in:

Selvam, C., Jebadass, R.J.J., Sundaram, D., & Shanmugam, L. A novel intuitionistic fuzzy generator for low-contrast color image enhancement technique. Information Fusion, 108 (2024), 102365. https://doi.org/10.1016/j.inffus.2024.102365

All mathematical formulations and algorithmic contributions belong to the original authors. This project provides an accessible, visual implementation for educational, research, and demonstration use.

Developed and adapted for general-purpose use by me.

License

This project is distributed under the MIT License. You are free to use, modify, and distribute the software, provided proper credit is given to the original authors and contributors.