Underwater Image Enhancement Based on Fourier-Fusion GAN

Underwater image enhancement is crucial for various applications, including marine ecology monitoring and underwater resource exploration. However, enhancing underwater images remains a challenging task due to issues such as low contrast, color distortion, and reduced visibility, which arise from light attenuation, scattering, and absorption, as well as the presence of suspended particles. These factors significantly degrade image quality and hinder accurate visual analysis. In this paper, we propose an Underwater Fourier-Fusion GAN (UFFGAN) for single underwater image enhancement. The proposed method leverages the Fast Fourier Transform (FFT) to convert images from the spatial domain to the frequency domain, enabling more effective enhancement of fine details. To address color distortion, Gaussian filters are applied to extract and refine low-frequency components for color correction. Finally, a Fusion GAN integrates the outputs from both enhancement processes, generating the final enhanced image. Experimental results demonstrate that the proposed approach effectively mitigates severe color cast and enhances image clarity, outperforming conventional methods in underwater image restoration.