GANID: a novel generative adversarial network for image dehazing

Manu, Chippy M; Sreeni, K. G.

doi:10.1007/s00371-022-02536-9

Cited by 15 publications

(2 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The quality assessment involves examining the generated images in comparison to real images or ground truth images. In the area of image-to-image generation, commonly addressed criteria include the Structural Similarity Index Measure (SSIM) and the Feature Similarity Index Measure (FSIM) [53][54][55][56]. SSIM is a widely used metric for assessing the structural similarity between two images.…”

Section: Test Comparisonsmentioning

confidence: 99%

Color Face Image Generation with Improved Generative Adversarial Networks

Chang,

Chung,

Chai

et al. 2024

Electronics

View full text Add to dashboard Cite

This paper focuses on the development of an improved Generative Adversarial Network (GAN) specifically designed for generating color portraits from sketches. The construction of the system involves using a GPU (Graphics Processing Unit) computing host as the primary unit for model training. The tasks that require high-performance calculations are handed over to the GPU host, while the user host only needs to perform simple image processing and use the model trained by the GPU host to generate images. This arrangement reduces the computer specification requirements for the user. This paper will conduct a comparative analysis of various types of generative networks which will serve as a reference point for the development of the proposed Generative Adversarial Network. The application part of the paper focuses on the practical implementation and utilization of the developed Generative Adversarial Network for the generation of multi-skin tone portraits. By constructing a face dataset specifically designed to incorporate information about ethnicity and skin color, this approach can overcome a limitation associated with traditional generation networks, which typically generate only a single skin color.

show abstract

Section: Test Comparisonsmentioning

confidence: 99%

Color Face Image Generation with Improved Generative Adversarial Networks

Chang,

Chung,

Chai

et al. 2024

Electronics

View full text Add to dashboard Cite

show abstract

“…This novel end-to-end design facilitates the integration of AOD-Net into other deep models (e.g., faster R-CNN) to enhance the performance of high-level tasks on blurry images. Manu suggested using a Generative Adversarial Network (GAN) for dehazing given blurry input images 33 . The proposed GAN architecture employs a Feature Residual Dense Network (FRDN) as the generator and a Markov discriminator with additional layers (PatchGAN) as the discriminator.…”

Section: Related Workmentioning

confidence: 99%

TMCD: Threshold multi-channel prior-based defogging algorithm for underground coal mine images

Fan,

Mao,

et al. 2024

Preprint

View full text Add to dashboard Cite

With the rapid development of image dehazing algorithms, the demand for effective dehazing solutions across various industries has markedly increased. However, the application effectiveness of most existing image dehazing algorithms within coal mine environments remains suboptimal. Consequently, this paper proposes an image dehazing algorithm based on a threshold multi-channel inspection method. The algorithm detects fog density using an enhanced color attenuation prior method, followed by image enhancement in fog-free areas and dehazing in foggy areas through threshold multi-channel inspection. During fog density detection, the algorithm incorporates texture information and illumination invariance features from the HSV space, enhancing adaptability and robustness to different lighting conditions. In the dehazing process, segregating foggy and fog-free images facilitates more accurate and reliable dehazing outcomes. Moreover, by constructing a multi-scale pyramid and employing a guided filtering approach, the algorithm achieves more precise estimation of the image transmittance, mitigates the blocky artifacts common in traditional methods. For video dehazing, a parameter reuse mechanism based on inter-frame similarity is designed, improving the real-time performance of video dehazing. The algorithm was tested on a coal mine dataset and on partial public datasets such as NH-Haze2 and Dense-Haze, achieving experimental results that surpass other algorithms.

show abstract