Underwater images typically suffer from severe color distortion, low contrast, and poor visibility due to light absorption and scattering in aquatic environments. To address these challenges, we propose a novel Dual-Attention GAN (DA-GAN), specifically designed to enhance contrast and correct color distortions in underwater images. DA-GAN features three key innovations: a Hybrid CNN-Attention Bottleneck Layer (HCABL) for efficient enhancement of local and global features, a Transformer-based Channel Attention Module (TCAM) for color distortion removal, and a Multi-Color Space Loss function to improve contrast and visibility. HCABL employs parallel attention mechanisms and depth-wise separable convolutions to enhance both global context and local spatial details. TCAM enhances the model’s sensitivity to various color channels, enabling effective correction of complex color distortions. Furthermore, the Multi-Color Space Loss function, tailored for the HSV color space and the image’s bright and dark channels, further enhance contrast and visibility. Comprehensive ablation studies validate the effectiveness of each component, while extensive comparison experiments demonstrate that DA-GAN achieves state-of-the-art or comparable performance across six public underwater image datasets (SUIM, UIEB, EUVP, LSUI, UCCS and SQUID). The source code is publicly available at https: //github.com/littlebearRedgood/DAGAN.

When using enhanced images for underwater object detection, issues such as detail loss and increased noise often arise, leading to decreased detection efficiency. To address this issue, we propose the Feature Fusion Network with Local Information Exchange (FFNLIE) for underwater object detection. We input raw and enhanced images into the Swin Transformer in parallel for feature extraction. Then, we propose a local information exchange module to enhance the feature extraction capability of the Swin Transformer. In order to fully utilize the complementary information of the two images, our feature fusion module consists of two core components: the Discrepancy Information Addition Block (DIAB) and the Common Information Addition Block (CIAB). The DIAB and CIAB are designed by utilizing and modifying cross-attention mechanisms, which can easily extract image discrepancy information and common information. Finally, the fused features are fed into the object detector to perform object detection tasks. The experimental findings demonstrate that the FFNLIE exhibited exceptional performance across four underwater datasets.