A rapid weighted median filter based on saliency region for underwater image denoising

In this paper, a novel weighted median filter model has been put forward with saliency detection for underwater image denoising. Saliency detection is first taken to only optionally highlight and prominent the salient region in the underwater images. One weighted median filter is further proposed to greatly accelerate the image denoising speed. It has been shown in our simulation experiments that the proposed method not only can provide better performance of suppressing noises and preserving the detailed features in underwater images, but also the running time can be largely shortened.