A learnable full-frequency transformer dual generative adversarial network for underwater image enhancement

Zheng, Shijian; Wang, Rujing; Zheng, Shitao; Wang, Liusan; Liu, Zhigui

doi:10.3389/fmars.2024.1321549

Cited by 4 publications

(1 citation statement)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Take underwater scenes as an example (depicted in Figure 1): they exhibit various degradation issues and diverse styles. By employing two restoration techniques, Learnable Full-frequency Transformer Dual Generative Adversarial Network (LFT-DGAN) [9], and Divide-and-Conquer network (DC-net) [10], higher-quality images can be generated. Each column represents the same scene but with varying detection outcomes using the cascaded RCNN detector, suggesting a potential link between underwater image restoration and object detection.…”

Section: Introductionmentioning

confidence: 99%

Underwater Fish Object Detection with Degraded Prior Knowledge

Zheng,

Wang,

Wang

2024

Electronics

Self Cite

View full text Add to dashboard Cite

Understanding fish distribution, behavior, and abundance is crucial for marine ecological research, fishery management, and environmental monitoring. However, the distinctive features of the underwater environment, including low visibility, light attenuation, water turbidity, and strong currents, significantly impact the quality of data gathered by underwater imaging systems, posing considerable challenges in accurately detecting fish objects. To address this challenge, our study proposes an innovative fish detection network based on prior knowledge of image degradation. In our research process, we first delved into the intrinsic relationship between visual image quality restoration and detection outcomes, elucidating the obstacles the underwater environment poses to object detection. Subsequently, we constructed a dataset optimized for object detection using image quality evaluation metrics. Building upon this foundation, we designed a fish object detection network that integrates a prompt-based degradation feature learning module and a two-stage training scheme, effectively incorporating prior knowledge of image degradation. To validate the efficacy of our approach, we develop a multi-scene Underwater Fish image Dataset (UFD2022). The experimental results demonstrate significant improvements of 2.4% and 2.5%, respectively, in the mAP index compared to the baseline methods ResNet50 and ResNetXT101. This outcome robustly confirms the effectiveness and superiority of our process in addressing the challenge of fish object detection in underwater environments.

show abstract

Section: Introductionmentioning

confidence: 99%

Underwater Fish Object Detection with Degraded Prior Knowledge

Zheng,

Wang,

Wang

2024

Electronics

Self Cite

View full text Add to dashboard Cite

show abstract

MSFFT-Net: A multi-scale feature fusion transformer network for underwater image enhancement

Wu,

Chen,

et al. 2025

Journal of Visual Communication and Image Representation

View full text Add to dashboard Cite

Underwater image restoration via attenuated incident optical model and background segmentation

Lin,

Sun,

2024

Front. Mar. Sci.

View full text Add to dashboard Cite

Underwater images typically exhibit low quality due to complex imaging environments, which impede the development of the Space-Air-Ground-Sea Integrated Network (SAGSIN). Existing physical models often ignore the light absorption and attenuation properties of water, making them incapable of resolving details and resulting in low contrast. To address this issue, we propose the attenuated incident optical model and combine it with a background segmentation technique for underwater image restoration. Specifically, we first utilize the features to distinguish the foreground region of the image from the background region. Subsequently, we introduce a background light layer to improve the underwater imaging model and account for the effects of non-uniform incident light. Afterward, we employ a new maximum reflection prior in the estimation of the background light layer to achieve restoration of the foreground region. Meanwhile, the contrast of the background region is enhanced by stretching the saturation and brightness components. Extensive experiments conducted on four underwater image datasets, using both classical and state-of-the-art (SOTA) algorithms, demonstrate that our method not only successfully restores textures and details but is also beneficial for processing images under non-uniform lighting conditions.

show abstract

A learnable full-frequency transformer dual generative adversarial network for underwater image enhancement

Cited by 4 publications

References 49 publications

Underwater Fish Object Detection with Degraded Prior Knowledge

Underwater Fish Object Detection with Degraded Prior Knowledge

MSFFT-Net: A multi-scale feature fusion transformer network for underwater image enhancement

Underwater image restoration via attenuated incident optical model and background segmentation

Contact Info

Product

Resources

About