An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement.
Research on underwater image processing has increased significantly in the past decade due to the precious resources that exist underwater. However, it is still a challenging problem to restore degraded underwater images. Existing prior-based methods show limited performance in many cases due to the...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2023-01-01
|
| Series: | PLoS ONE |
| Online Access: | https://doi.org/10.1371/journal.pone.0279945 |
| _version_ | 1797866782726291456 |
|---|---|
| author | Zeru Lan Bin Zhou Weiwei Zhao Shaoqing Wang |
| author_facet | Zeru Lan Bin Zhou Weiwei Zhao Shaoqing Wang |
| author_sort | Zeru Lan |
| collection | DOAJ |
| description | Research on underwater image processing has increased significantly in the past decade due to the precious resources that exist underwater. However, it is still a challenging problem to restore degraded underwater images. Existing prior-based methods show limited performance in many cases due to their reliance on hand-crafted features. Therefore, in this paper, we propose an effective unsupervised generative adversarial network(GAN) for underwater image restoration. Specifically, we embed the idea of contrastive learning into the model. The method encourages two elements (corresponding patches) to map the similar points in the learned feature space relative to other elements (other patches) in the data set, and maximizes the mutual information between input and output through PatchNCE loss. We design a query attention (Que-Attn) module, which compares feature distances in the source domain, and gives an attention matrix and probability distribution for each row. We then select queries based on their importance measure calculated from the distribution. We also verify its generalization performance on several benchmark datasets. Experiments and comparison with the state-of-the-art methods show that our model outperforms others. |
| first_indexed | 2024-04-09T23:30:48Z |
| format | Article |
| id | doaj.art-7932b34529834d2a93c9cbf972980586 |
| institution | Directory Open Access Journal |
| issn | 1932-6203 |
| language | English |
| last_indexed | 2024-04-09T23:30:48Z |
| publishDate | 2023-01-01 |
| publisher | Public Library of Science (PLoS) |
| record_format | Article |
| series | PLoS ONE |
| spelling | doaj.art-7932b34529834d2a93c9cbf9729805862023-03-21T05:31:49ZengPublic Library of Science (PLoS)PLoS ONE1932-62032023-01-01181e027994510.1371/journal.pone.0279945An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement.Zeru LanBin ZhouWeiwei ZhaoShaoqing WangResearch on underwater image processing has increased significantly in the past decade due to the precious resources that exist underwater. However, it is still a challenging problem to restore degraded underwater images. Existing prior-based methods show limited performance in many cases due to their reliance on hand-crafted features. Therefore, in this paper, we propose an effective unsupervised generative adversarial network(GAN) for underwater image restoration. Specifically, we embed the idea of contrastive learning into the model. The method encourages two elements (corresponding patches) to map the similar points in the learned feature space relative to other elements (other patches) in the data set, and maximizes the mutual information between input and output through PatchNCE loss. We design a query attention (Que-Attn) module, which compares feature distances in the source domain, and gives an attention matrix and probability distribution for each row. We then select queries based on their importance measure calculated from the distribution. We also verify its generalization performance on several benchmark datasets. Experiments and comparison with the state-of-the-art methods show that our model outperforms others.https://doi.org/10.1371/journal.pone.0279945 |
| spellingShingle | Zeru Lan Bin Zhou Weiwei Zhao Shaoqing Wang An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement. PLoS ONE |
| title | An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement. |
| title_full | An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement. |
| title_fullStr | An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement. |
| title_full_unstemmed | An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement. |
| title_short | An optimized GAN method based on the Que-Attn and contrastive learning for underwater image enhancement. |
| title_sort | optimized gan method based on the que attn and contrastive learning for underwater image enhancement |
| url | https://doi.org/10.1371/journal.pone.0279945 |
| work_keys_str_mv | AT zerulan anoptimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement AT binzhou anoptimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement AT weiweizhao anoptimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement AT shaoqingwang anoptimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement AT zerulan optimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement AT binzhou optimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement AT weiweizhao optimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement AT shaoqingwang optimizedganmethodbasedonthequeattnandcontrastivelearningforunderwaterimageenhancement |