PRAGAN: Progressive Recurrent Attention GAN with Pretrained ViT Discriminator for Single-Image Deraining

Wei, Bingcai; Wang, Di; Wang, Zhuang; Zhang, Liye

doi:10.3390/s22249587

Cited by 3 publications

(2 citation statements)

References 65 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite the growing interest in GANs for data generation, there are relatively few methods in the field of rain fields synthesis [14,34,35]. Existing work mainly focuses on singleimage Derain, using GAN learning to generate multiple pairs of rainy-clear images, or adopting multi-scale cascade methods to analyse rainy images and generate rain maps [31][32][33]. The codec structure has also been introduced into the image rain removal task [36].…”

Section: Related Workmentioning

confidence: 99%

Attention‐generative adversarial networks for simulating rain field

Li,

Zhao,

et al. 2024

IET Image Processing

View full text Add to dashboard Cite

The synthesis of rain fields is essential in multiple research fields and applications, including Single‐image Derain. However, there is a lack of research on simulated rain fields, and the existing rain field generation models struggle to capture complex spatial distributions and generate truly random rain fields. To address this, the authors propose a generative adversarial networks‐based rain field generation network, which consists of a generator, a discriminator, and a feature extraction block that can produce realistic and complex rain fields. The authors’ experiments demonstrate that this method achieves an average Frechet Inception Distance score of 0.035, and user studies indicate that the generated rain distribution looks naturally.

show abstract

Section: Related Workmentioning

confidence: 99%

Attention‐generative adversarial networks for simulating rain field

Li,

Zhao,

et al. 2024

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…Li et al [12] used a recalibration network to progressively remove rain streaks at different stages and obtain a clean background image. Zhang et al [13] applied the GAN [14,15] to image deraining, and used an ensemble residual perceptual classifier to adapt to the rainwater density information. Although the performance of deep learning algorithms has significantly improved compared to traditional algorithms, there are still some issues, such as the size and direction of the rain streaks being ignored, resulting in residual rain; during the rain removal process, due to the inability to distinguish between rain streaks and background textures, the background details are lost.…”

Section: Introductionmentioning

confidence: 99%

Symmetric Enhancement of Visual Clarity through a Multi-Scale Dilated Residual Recurrent Network Approach for Image Deraining

2023

View full text Add to dashboard Cite

Images captured during rainy days present the challenge of maintaining a symmetrical balance between foreground elements (like rain streaks) and the background scenery. The interplay between these rain-obscured images is reminiscent of the principle of symmetry, where one element, the rain streak, overshadows or disrupts the visual quality of the entire image. The challenge lies not just in eradicating the rain streaks but in ensuring the background is symmetrically restored to its original clarity. Recently, numerous deraining algorithms that employ deep learning techniques have been proposed, demonstrating promising results. Yet, achieving a perfect symmetrical balance by effectively removing rain streaks from a diverse set of images, while also symmetrically restoring the background details, is a monumental task. To address this issue, we introduce an image-deraining algorithm that leverages multi-scale dilated residual recurrent networks. The algorithm begins by utilizing convolutional activation layers to symmetrically process both the foreground and background features. Then, to ensure the symmetrical dissemination of the characteristics of rain streaks and the background, it employs long short-term memory networks in conjunction with gated recurrent units across various stages. The algorithm then incorporates dilated residual blocks (DRB), composed of dilated convolutions with three distinct dilation factors. This integration expands the receptive field, facilitating the extraction of deep, multi-scale features of both the rain streaks and background information. Furthermore, considering the complex and diverse nature of rain streaks, a channel attention (CA) mechanism is incorporated to capture richer image features and enhance the model’s performance. Ultimately, convolutional layers are employed to fuse the image features, resulting in a derained image. An evaluation encompassing seven benchmark datasets, assessed using five quality metrics against various conventional and modern algorithms, confirms the robustness and flexibility of our approach.

show abstract

The visual motion blur elimination method for silicon nitride bearing roller fissures based on U-Net asymmetric multi-scale feature fusion

Deng,

Li,

Yang

et al. 2024

AIP Advances

View full text Add to dashboard Cite

The visual motion blur imaging for the feature recognition process of silicon nitride bearing roller fissures is a pathological problem. This is solved by proposing squeeze-and-excitation asymmetric fusion of multi-scale features with high-frequency loss attention coupled U-Net (MHU-Net). The visual motion blur elimination of fissure features on silicon nitride bearing rollers is achieved. In the deblurring model, the multi-scale feature information on silicon nitride bearing roller fissures is blocked and there is weak correlation between channels. A design for an asymmetric fusion multi-scale feature module under the channel information compression–excitation mode is proposed. It successfully balances the channel information from different scales while integrating multi-scale features in image fusion. The high-frequency region of fissure features on silicon nitride bearing rollers is analyzed. Around the high-frequency feature loss in the multi-frequency domain of images combined with spatial feature loss, a multi-frequency band high-frequency loss attention module is built. Then, the complete structural details of silicon nitride bearing roller fissures are obtained. The proposed algorithm achieves a peak signal-to-noise ratio of 27.58 and a structural similarity of 0.847 on our self-made silicon nitride defect motion dataset. The visual motion blur of fissure features is noticeably eliminated. The restored image exhibits complete details in the feature structures and overall region smoothness.

show abstract

PRAGAN: Progressive Recurrent Attention GAN with Pretrained ViT Discriminator for Single-Image Deraining

Cited by 3 publications

References 65 publications

Attention‐generative adversarial networks for simulating rain field

Attention‐generative adversarial networks for simulating rain field

Symmetric Enhancement of Visual Clarity through a Multi-Scale Dilated Residual Recurrent Network Approach for Image Deraining

The visual motion blur elimination method for silicon nitride bearing roller fissures based on U-Net asymmetric multi-scale feature fusion

Contact Info

Product

Resources

About