Double-U Net: A Novel Approach for Deep Learning Based Image Defogging


  • Anu Bajaj
  • Ankit Bhardwaj
  • Yessica Tuteja
  • Mayank Jindal
  • Surbhi
  • Rahul Saini
  • Ajith Abraham


Image defogging has become a major difficulty in the quickly developing field of digital photography. Addressing this issue is crucial, given the growing need for crisp, high-quality images in industries like social networking, entertainment, navigation, and surveillance. Researchers from all over the world have put out a variety of presumptions and techniques to improve clarity in hazy or foggy images. This research paper proposed a new deep learning algorithm, the Double U-Net algorithm. Which is a concatenation of two similar or distinct U Net architectures to produce the best possible outcomes while enhancing visibility in cloudy photos. We conduct a thorough analysis to compare the effectiveness of the proposed algorithm with other state-of-the-art defogging methods, considering factors such as robustness to varying fog intensities and image features, computing efficiency, and visual quality. It is observed that the proposed architecture outperformed other techniques in terms of PSNR (26.88) and SSIM (0.99979). The findings demonstrate that the proposed algorithm performs exceptionally well in improving visibility and recovering fine-grained image information under various atmospheric situations.


Download data is not yet available.




How to Cite

Anu Bajaj, Ankit Bhardwaj, Yessica Tuteja, Mayank Jindal, Surbhi, Rahul Saini, & Ajith Abraham. (2024). Double-U Net: A Novel Approach for Deep Learning Based Image Defogging . International Journal of Computer Information Systems and Industrial Management Applications, 16(3), 12. Retrieved from



Original Articles