Retracted: Image Enhancement Using Convolutional Neural Networks

Abstract

This paper presented a Convolutional Neutral Network (CNN) based technique to perform image enhancement the contrast of images can be adaptively improved. At the point when an image is caught under inadequate dimensions, the pixel values are in a low dynamic range, which will cause image quality to descend apparently. Hence, it is important to improve the nature of images. So as to safeguard image surfaces however much as could reasonably be expected, and the use SSIM loss. DnCNN grasps the advancement in profound engineering, learning calculation and regularization technique into image denoising. DnCNN certainly evacuates the inactive clean image in the shrouded unit layers. Finally, this CNN method has increased the PSNR value and reduce the SSIM value.