SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal
Deraining of images plays a pivotal role in computer vision by addressing the challenges posed by rain, enhancing visibility, and refining image quality by eliminating rain streaks. Traditional methods often fall short of effectively handling intricate rain patterns, resulting in incomplete removal....
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10466540/ |
| _version_ | 1797236588426559488 |
|---|---|
| author | Maheshkumar H. Kolekar Samprit Bose Abhishek Pai |
| author_facet | Maheshkumar H. Kolekar Samprit Bose Abhishek Pai |
| author_sort | Maheshkumar H. Kolekar |
| collection | DOAJ |
| description | Deraining of images plays a pivotal role in computer vision by addressing the challenges posed by rain, enhancing visibility, and refining image quality by eliminating rain streaks. Traditional methods often fall short of effectively handling intricate rain patterns, resulting in incomplete removal. In this paper, we propose an innovative deep learning-based deraining model leveraging a modified residual UNet and a multiscale attention-guided convolutional neural network module as a discriminator within a conditional generative adversarial network framework. The proposed approach introduces custom hyperparameters and a tailored loss function to facilitate the efficient removal of rain streaks from images. Evaluation on both synthetic and real-world datasets showcases superior performance, as indicated by improved image evaluation metrics such as PSNR, SSIM, and NIQE. The effectiveness of our model extends to improving both rainy and foggy images. We also conducted a comparative analysis of computational complexity in terms of running time, GFLOPs, and no. of parameters against other state-of-the-art methods to demonstrate our model’s superiority. |
| first_indexed | 2024-04-24T17:06:14Z |
| format | Article |
| id | doaj.art-80b12746dbfa4ac68ba06bddb56def20 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-24T17:06:14Z |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-80b12746dbfa4ac68ba06bddb56def202024-03-28T23:00:26ZengIEEEIEEE Access2169-35362024-01-0112438744388810.1109/ACCESS.2024.337590910466540SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak RemovalMaheshkumar H. Kolekar0https://orcid.org/0000-0002-4272-3528Samprit Bose1https://orcid.org/0009-0008-0045-9475Abhishek Pai2https://orcid.org/0009-0008-3201-1181Department of Electrical Engineering, Indian Institute of Technology Patna, Patna, Bihar, IndiaDepartment of Electrical Engineering, Indian Institute of Technology Patna, Patna, Bihar, IndiaDepartment of Information Technology, Bharatiya Vidya Bhavan's Sardar Patel Institute of Technology, Mumbai, Maharashtra, IndiaDeraining of images plays a pivotal role in computer vision by addressing the challenges posed by rain, enhancing visibility, and refining image quality by eliminating rain streaks. Traditional methods often fall short of effectively handling intricate rain patterns, resulting in incomplete removal. In this paper, we propose an innovative deep learning-based deraining model leveraging a modified residual UNet and a multiscale attention-guided convolutional neural network module as a discriminator within a conditional generative adversarial network framework. The proposed approach introduces custom hyperparameters and a tailored loss function to facilitate the efficient removal of rain streaks from images. Evaluation on both synthetic and real-world datasets showcases superior performance, as indicated by improved image evaluation metrics such as PSNR, SSIM, and NIQE. The effectiveness of our model extends to improving both rainy and foggy images. We also conducted a comparative analysis of computational complexity in terms of running time, GFLOPs, and no. of parameters against other state-of-the-art methods to demonstrate our model’s superiority.https://ieeexplore.ieee.org/document/10466540/Image derainingdeep learningresidual UNetfoggy image enhancement |
| spellingShingle | Maheshkumar H. Kolekar Samprit Bose Abhishek Pai SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal IEEE Access Image deraining deep learning residual UNet foggy image enhancement |
| title | SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal |
| title_full | SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal |
| title_fullStr | SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal |
| title_full_unstemmed | SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal |
| title_short | SARain-GAN: Spatial Attention Residual UNet Based Conditional Generative Adversarial Network for Rain Streak Removal |
| title_sort | sarain gan spatial attention residual unet based conditional generative adversarial network for rain streak removal |
| topic | Image deraining deep learning residual UNet foggy image enhancement |
| url | https://ieeexplore.ieee.org/document/10466540/ |
| work_keys_str_mv | AT maheshkumarhkolekar sarainganspatialattentionresidualunetbasedconditionalgenerativeadversarialnetworkforrainstreakremoval AT sampritbose sarainganspatialattentionresidualunetbasedconditionalgenerativeadversarialnetworkforrainstreakremoval AT abhishekpai sarainganspatialattentionresidualunetbasedconditionalgenerativeadversarialnetworkforrainstreakremoval |