Enhancing Underwater Imagery Clarity: A Novel Approach with Zero DCE NET

Underwater imaging poses different challenges to the enhancement techniques because of the distortion, quality, absence of natural light and many other factors. In this paper, we have explored the efficiency of deep learning based algorithms in the realm of image enhancement. Our methodology explores the implementation and optimization of Zero reference Deep Curve Estimation Network to refine the enhancement process. It can be observed that the tailored network significantly improve the visual quality of underwater images by preserving its spatial features. Results are evaluated through both quantitative and qualitative measures which reveal substantial improvement of traditional algorithms. Different loss functions along with the evaluation metrics are used to underscore the potential of the proposed algorithm.