Transforming Images with GANs: Dehazing and Edge Detection for Enhanced Features

DOI

10.2991/978-94-6463-866-0_84How to use a DOI?

Keywords

Image dehazing; edge detection; K-means clustering; CNN; GAN

Abstract

Image restoration and enhancement have significantly progressed in restoring degraded images with haze, missing areas, and structural ill-clarity. Conventional restoration techniques are not able to retain global structures and fine textures simultaneously, resulting in artifacts and loss of quality. In an endeavour to overcome the above-mentioned shortcomings, the present work discusses a novel dual-stage GAN-driven structure and texture-aware image restoration model incorporating dehazing and edge enhancement modules. The proposed model has two stages, including a structure-aware generator and a texture refinement generator to restore real structure and texture, followed by a GAN-based dehazing module to enhance image brightness in poor light conditions. K-means clustering is used to perform efficient feature segmentation, further enhancing subsequent edge detection using the Holistically Nested Edge Detection (HED) algorithm. Diversified benchmark datasets like RESIDE, BSD500, COCO, and Places2-MIT are used to obtain input images. Dehazed and corrupt images are subjected to processing with clustering to remove noise, thus allowing accurate edge recovery and structure prediction. Principles of membrane computing, like hierarchical structuring and compartmentalized learning strategies, are applied to optimize the recovery of structures and textures efficiently. Feature patch discriminators are utilized to detect real and generated patches during adversarial training. According to experiments, the suggested approach produces an SSIM value of 0.948, a low computational time of 0.045 seconds, and an excellent PSNR value of 35.7 dB. The proposed methodology efficiently restores degraded images with improved structure retention, texture reality, and enhanced feature clarity compared to existing techniques.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - S. Malathy
AU  - N. Bharathi
AU  - S. Prasanna Devi
PY  - 2025
DA  - 2025/10/31
TI  - Transforming Images with GANs: Dehazing and Edge Detection for Enhanced Features
BT  - Proceedings of the International Conference on Intelligent Systems and Digital Transformation (ICISD 2025)
PB  - Atlantis Press
SP  - 1050
EP  - 1066
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6463-866-0_84
DO  - 10.2991/978-94-6463-866-0_84
ID  - Malathy2025
ER  -