<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>The traditional rain removal algorithm needs to optimize a large number of parameters, and it is only effective for rain of a specific shape, and the model generalization ability is poor. In recent years, the performance of rain removal methods based on deep learning is better than many traditional methods, but there are problems such as incomplete or excessive rain removal, and incomplete texture reconstruction of background details. This paper proposes a rain removal network based on generative confrontation, which connects the high and low frequency parts and integrates them into the model. At the same time, the attention mechanism cyclic neural network is organically combined, which can better preserve the background texture while removing rain. Theoretical can produce better rain streak removal with better color distortion.</p>
</abstract>

The traditional rain removal algorithm needs to optimize a large number of parameters, and it is only effective for rain of a specific shape, and the model generalization ability is poor. In recent years, the performance of rain removal methods based on deep learning is better than many traditional methods, but there are problems such as incomplete or excessive rain removal, and incomplete texture reconstruction of background details. This paper proposes a rain removal network based on generative confrontation, which connects the high and low frequency parts and integrates them into the model. At the same time, the attention mechanism cyclic neural network is organically combined, which can better preserve the background texture while removing rain. Theoretical can produce better rain streak removal with better color distortion.

Research and Implementation of Image Rain Removal Based on Deep Learning

School of Computer Science and Engineering

International Journal of Advanced Network, Monitoring and Controls

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

{"article-title":"Research and Implementation of Image Rain Removal Based on Deep Learning"}

The traditional rain removal algorithm needs to optimize a large number of parameters, and it is only effective for rain of a specific shape, and the model...

Datasets	GMM	JORDER	RESCAN	Ours
PSNR	25.9725	26.4012	28.7863	29.1035
SSIM	0.9181	0.9246	0.9355	0.9661

Research and Implementation of Image Rain Removal Based on Deep Learning

Data publikacji: 24 maj 2023

Zakres stron: 25 - 32

DOI: https://doi.org/10.2478/ijanmc-2022-0024

Słowa kluczowe
Deep Learning, Generative Confrontation, High and Low Frequency, Attention Mechanism, Background Texture

© 2022 Dong Wang et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Comparison of experimental results on synthetic datasets

Research and Implementation of Image Rain Removal Based on Deep Learning

Data publikacji: 24 maj 2023

Zakres stron: 25 - 32

DOI: https://doi.org/10.2478/ijanmc-2022-0024

Słowa kluczoweDeep Learning, Generative Confrontation, High and Low Frequency, Attention Mechanism, Background Texture

© 2022 Dong Wang et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Comparison of experimental results on synthetic datasets

Słowa kluczowe
Deep Learning, Generative Confrontation, High and Low Frequency, Attention Mechanism, Background Texture