Underwater Image Enhancement Using MIRNet

Abstract

In recent years, enhancement of underwater images is a challenging task, which is gaining priority since the human eye cannot perceive images under water. The significant details underwater are not clearly captured using the conventional image acquisition techniques, and also they are expensive. Hence, the quality of the image processing algorithms can be enhanced in the absence of costly and reliable acquisition techniques. Traditional algorithms have certain limitations in the case of these images with varying degrees of fuzziness and color deviation. In the proposed model, the authors used a deep learning model for underwater image enhancement. First, the original image is pre-processed by the white balance algorithm for colour correction and the contrast of the image is improved using the contrast enhancement technique. Next, the pre-processed image is given to the MIRNet for enhancement. MIRNet is a deep learning framework that can be used to enhance the low-light level images. The enhanced image quality is measured using peak signal-to-noise ratio (PSNR), root mean square error (RMSE), and structural similarity index (SSIM) parameters.