Underwater Image Enhancement Using White Balancing with Color Correction Based on Recurrent Neural Network

Underwater imaging presents unique challenges due to light absorption, scattering, and color distortion. Conventional image enhancement methods often fail to address these issues adequately, resulting in degraded image quality. In this study, we propose a novel approach for underwater image enhancement and white balancing with color compensation using recurrent neural networks (RNNs). Our method utilizes the temporal dependencies in sequential underwater image data to effectively capture and learn complex patterns of color distortion and contrast variation. Specifically, we employ a recurrent neural network architecture, specifically long short-term memory (LSTM) networks, to model the sequential nature of underwater images and adaptively adjust color balance and compensation parameters. To evaluate the efficacy of our approach, we conduct experiments on a diverse dataset of underwater images captured in various underwater environments. Quantitative and qualitative assessments demonstrate the superior performance of our method in enhancing color fidelity, improving contrast, and restoring details in underwater images compared to traditional techniques. Furthermore, our approach exhibits robustness in handling challenging conditions such as low visibility and fluctuating lighting conditions encountered in underwater settings. These results underscore the potential of recurrent neural networks for addressing the inherent complexities of underwater image processing and advancing underwater photography and exploration