• Snehal Teli
• Rupali Shelke

Due to environmental scattering, images taken in the underwater environment still exhibit color distortion, loss of resolution, and decreased contrast. The paper presents a method for enhancing the aesthetic appeal of underwater photography. Yet, because of unwanted staining, decreased contrast, and detail loss brought on by light scattering and absorption, photos that are directly taken in the marine environment are still highly damaged, drastically limiting the amount of information that can be extracted from the image. This study introduces the U-Net, a CNN-based network, and provides an end-to-end framework for subaquatic picture improvement to address this problem. The two workouts that are utilized to train the U-net are colour correction and haze removal. It is possible to simultaneously acquire a potent feature representation for both tasks using this dual training strategy. To manage the training of U-net, we generate 1000 synthetic training images based on the physical underwater imaging model. Depending on the language and framework used.

Machine Learning, CNN algorithm, Feature Extraction, Image Classification.

"Underwater Image Enhancements Using CNN And U-Net", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 7, page no.d243-d248, July-2023, Available :http://www.jetir.org/papers/JETIR2307331.pdf

"Underwater Image Enhancements Using CNN And U-Net", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 7, page no. ppd243-d248, July-2023, Available at : http://www.jetir.org/papers/JETIR2307331.pdf