What question did this study set out to answer?

The aim is to enhance underwater images by integrating deep learning and optimization techniques.

Underwater low-light image enhancement using a hybrid CNN–transformer framework with multi-scale feature fusion and cheetah optimization

View Full Paper

Ask AI

Bookmark

Authors

A hybrid framework improves image clarity in underwater environments, highlighting the potential for better AUV navigation.

The aim is to enhance underwater images by integrating deep learning and optimization techniques.
Developed a hybrid CNN-transformer architecture for feature extraction and modeling.
Implemented Cheetah Optimization Algorithm to enhance parameter tuning during training.
Utilized the Multi-Residual Refinement Plus Plus module for iterative image improvement using residual learning.
Trained and evaluated the model on the UIEB benchmark underwater dataset.
The proposed method consistently outperformed state-of-the-art image enhancement techniques.
Achieved improved quantitative metrics, including PSNR, SSIM, UIQM, and UIQI.
Demonstrated superior visual quality in a variety of challenging underwater scenarios.

Midhunchakkaravarthy et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark