Underwater image enhancement based on zero-shot learning and level adjustment

Abstract

Light is scattered and partially absorbed while traveling through water, hence, underwater captured images often exhibit issues such as low contrast, detail blurring, color attenuation, and low illumination. To improve the visual performance of underwater imaging, herein, we propose a two-step method of zero-shot dehazing and level adjustment. In the newly developed approach, the original image is fed into a “zero-shot” dehazing network and further enhanced by an improved level adjustment methodology combined with auto-contrast. By conducting experiments, we then compare the performance of the proposed method with six classical state-of-the-art methods. The qualitative results confirm that the proposed method is capable of effectively removing haze, correcting color deviations, and maintaining the naturalness of images. We further perform a quantitative evaluation, revealing that the proposed method outperforms the comparison methods in terms of peak signal-to-noise ratio and structural similarity. The enhancement results are also measured by employing the underwater color image quality evaluation index (UCIQE), indicating that the proposed approach exhibits the highest mean values of 0.58 and 0.53 on the two data sets. The experimental results collectively validate the efficiency of the proposed methodology in enhancing underwater blurred images.