Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{184390,
        author = {Neetin Kumar and Bhoomika sahu},
        title = {Transforming Underwater Visuals with Deep Learning for Enhanced Observation and Analysis},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {4},
        pages = {1281-1288},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=184390},
        abstract = {The improvement of underwater images is essential for the valid visual observation, examination, and analyzing them in underwater scenes. Nevertheless, underwater image quality is usually degraded due to light attenuation, scattering, turbidity and noise, leading to visibility degradation, color distortion, and structural detail loss. Histogram equalization (HE) and contrast-limited adaptive histogram equalization (CLAHE) are classical ways to enhance the contrast, but suffer from limited success when dealing with the diversity and high variability of underwater scenes. To solve these problems, the study presents a novel deep learning framework of combining CNNs, attention mechanism, and feature fusion for effective underwater image restoration. The model can be trained in both paired and unpaired training, and with data augmentation, normalization and other water quality parameters such as depth, salinity, and turbidity. Experimental results on benchmark data sets (EUVP, UIEB, and U45) demonstrate that the proposed method outperforms existing classical and state-of-the-art methods with a peak signal-to-noise ratio (PSNR) of 32.8 dB, a structural similarity index (SSIM) of 0.92, and the lowest CIEDE2000 value of 11.5. The proposed work provides a scalable platform for underwater exploration, marine biology, environmental monitoring and robotic.},
        keywords = {Underwater Image Enhancement, Deep Learning, Convolutional Neural Networks (CNNs), Attention Mechanism, Feature Fusion, Marine Imaging.},
        month = {September},
        }