Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image Dehazing

Fog reduces image contrast and clarity, creating challenges for applications such as autonomous driving and remote sensing. This study proposes a series of CycleGAN modifications for single image dehazing using unpaired data, integrating residual blocks, attention mechanisms, VGG19-based perceptual...

Full description

Saved in:
Bibliographic Details
Main Authors: Sani Moch Sopian, Arief Suryadi Satyawan, Mokhammad Mirza Etnisa Haqiqi, Helfy Susilawati, Beni Wijaya, Khaulyca Arva Artemysia, Firman, Muhammad Ikbal Samie
Format: Article
Language:English
Published: Center for Research and Community Service, Institut Informatika Indonesia Surabaya 2025-07-01
Series:Teknika
Subjects:
Online Access:https://ejournal.ikado.ac.id/index.php/teknika/article/view/1235
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849429554109612032
author Sani Moch Sopian
Arief Suryadi Satyawan
Mokhammad Mirza Etnisa Haqiqi
Helfy Susilawati
Beni Wijaya
Khaulyca Arva Artemysia
Firman
Muhammad Ikbal Samie
author_facet Sani Moch Sopian
Arief Suryadi Satyawan
Mokhammad Mirza Etnisa Haqiqi
Helfy Susilawati
Beni Wijaya
Khaulyca Arva Artemysia
Firman
Muhammad Ikbal Samie
author_sort Sani Moch Sopian
collection DOAJ
description Fog reduces image contrast and clarity, creating challenges for applications such as autonomous driving and remote sensing. This study proposes a series of CycleGAN modifications for single image dehazing using unpaired data, integrating residual blocks, attention mechanisms, VGG19-based perceptual loss, and haze-aware loss. Among ten architectural variants, Modification 10 combining perceptual and haze-aware loss achieved the best overall performance. Quantitatively, it showed stable generator losses (0.91 for Gen G, 0.57 for Gen F), with improved discriminator performance (Disc X: 0.59, Disc Y: 0.47), indicating better training stability and image realism. Additionally, it offered competitive PSNR (7.99), strong SSIM (0.4202), and low LPIPS (0.6577), confirming its effectiveness in both pixel-level accuracy and perceptual quality. Qualitatively, this model generated clearer, more natural images with improved edge sharpness and detail preservation. These findings demonstrate that the modified CycleGAN significantly enhances dehazing performance and presents a valuable contribution to deep learning-based image restoration.
format Article
id doaj-art-d33833b75c6344d58e2dbc2d7249ed4b
institution Kabale University
issn 2549-8037
2549-8045
language English
publishDate 2025-07-01
publisher Center for Research and Community Service, Institut Informatika Indonesia Surabaya
record_format Article
series Teknika
spelling doaj-art-d33833b75c6344d58e2dbc2d7249ed4b2025-08-20T03:28:19ZengCenter for Research and Community Service, Institut Informatika Indonesia SurabayaTeknika2549-80372549-80452025-07-0114210.34148/teknika.v14i2.1235Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image DehazingSani Moch SopianArief Suryadi SatyawanMokhammad Mirza Etnisa HaqiqiHelfy SusilawatiBeni WijayaKhaulyca Arva ArtemysiaFirmanMuhammad Ikbal Samie Fog reduces image contrast and clarity, creating challenges for applications such as autonomous driving and remote sensing. This study proposes a series of CycleGAN modifications for single image dehazing using unpaired data, integrating residual blocks, attention mechanisms, VGG19-based perceptual loss, and haze-aware loss. Among ten architectural variants, Modification 10 combining perceptual and haze-aware loss achieved the best overall performance. Quantitatively, it showed stable generator losses (0.91 for Gen G, 0.57 for Gen F), with improved discriminator performance (Disc X: 0.59, Disc Y: 0.47), indicating better training stability and image realism. Additionally, it offered competitive PSNR (7.99), strong SSIM (0.4202), and low LPIPS (0.6577), confirming its effectiveness in both pixel-level accuracy and perceptual quality. Qualitatively, this model generated clearer, more natural images with improved edge sharpness and detail preservation. These findings demonstrate that the modified CycleGAN significantly enhances dehazing performance and presents a valuable contribution to deep learning-based image restoration. https://ejournal.ikado.ac.id/index.php/teknika/article/view/1235DehazingCycleGANDeep LearningResidual BlockAttention Mechanism
spellingShingle Sani Moch Sopian
Arief Suryadi Satyawan
Mokhammad Mirza Etnisa Haqiqi
Helfy Susilawati
Beni Wijaya
Khaulyca Arva Artemysia
Firman
Muhammad Ikbal Samie
Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image Dehazing
Teknika
Dehazing
CycleGAN
Deep Learning
Residual Block
Attention Mechanism
title Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image Dehazing
title_full Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image Dehazing
title_fullStr Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image Dehazing
title_full_unstemmed Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image Dehazing
title_short Development of a Modified CycleGAN Model with Residual Blocks and Perceptual Loss for Image Dehazing
title_sort development of a modified cyclegan model with residual blocks and perceptual loss for image dehazing
topic Dehazing
CycleGAN
Deep Learning
Residual Block
Attention Mechanism
url https://ejournal.ikado.ac.id/index.php/teknika/article/view/1235
work_keys_str_mv AT sanimochsopian developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing
AT ariefsuryadisatyawan developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing
AT mokhammadmirzaetnisahaqiqi developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing
AT helfysusilawati developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing
AT beniwijaya developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing
AT khaulycaarvaartemysia developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing
AT firman developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing
AT muhammadikbalsamie developmentofamodifiedcycleganmodelwithresidualblocksandperceptuallossforimagedehazing