v1.0.0: Initial Release & Research Assets

What's Included

Release Assets

• report.pdf: Complete research report covering dehazing principles, algorithm implementations, parameter tuning, and experimental comparative evaluations.
• ImagesForTesting.zip: Curated benchmark test dataset containing diverse real-world and synthetic hazy images for testing and replication.

Core Features in v1.0.0

• 6 Dehazing Algorithms:
  • Dark Channel Prior (DCP)
  • Guided Image Filter DCP
  • Retinex (Single-Scale SSR & Multi-Scale MSR)
  • Contrast Limited Adaptive Histogram Equalization (CLAHE + USM)
  • Three-Region Partitioned Improved DCP (Sky / Bright / General)
  • Paper-Improved Adaptive Dark Channel Dehazing
• Dual Evaluation Metric Framework (9 Metrics):
  • Standard Distortion: PSNR, SSIM, MSE, and processing latency (ms).
  • Specialized Dehazing Assessment: Grayscale contrast standard deviation ($\sigma$), HSI hue fidelity ($d$), valid gradient details ($I_{\text{valid}}$), structural component ($S$), and unified composite score ($q$).
• Modern Desktop GUI (PySide6):
  • Dynamic algorithm registry with auto-generated parameter controls.
  • Asynchronous background calculation (QThread) with sequence invalidation.
  • Side-by-side comparison, drag-and-drop loading, timestamped undo history, and live console logging.

---

Third-Party Asset Attribution & Copyright Notice

The images packaged in ImagesForTesting.zip are not original works of this project. They were collected from public free-stock image platforms and are redistributed here solely for academic research, algorithm evaluation, and experimental replication. All copyright in these images remains with their original authors / photographers, and no ownership whatsoever is claimed by the repository owner or contributors.

• Source platforms: Pixabay and Pexels.
• License scope: each image is used under the license of its source platform (Pixabay Content License / Pexels License), which permits free use but does not transfer copyright.
• Purpose limitation: the images are included as part of a research artifact for testing a dehazing algorithm. They are not being redistributed as stock photography or as a standalone image collection.
• Credits: the original authors are indicated by the file names themselves.

File	Source platform
mostafa_meraji-istanbul-4917452_640.jpg	Pixabay
mostafa_meraji-istanbul-city-4923061_640.jpg	Pixabay
mostafa_meraji-istanbul-city-4923064_640.jpg	Pixabay
pexels-ahrphotography-29751756.jpg	Pexels
pexels-allen-boguslavsky-1344061-36431775.jpg	Pexels
pexels-helen1-30455399.jpg	Pexels
pexels-shuaizhi-tian-485596-30376943.jpg	Pexels
pexels-suki-lee-110686949-18672898.jpg	Pexels
pexels-suki-lee-110686949-18685047.jpg	Pexels
pexels-unlime-8272122.jpg	Pexels
pexels-yuraforrat-15327226.jpg	Pexels

report.pdf is the original work of the repository author.

Takedown request: if you are the copyright holder of any asset listed above and do not wish for it to be included in this release, please open an issue in this repository and the asset will be removed promptly.