Deep CNN denoiser prior for blurred images restoration with multiplicative noise

Multiplicative noise always appears in the formation of image of synthetic aperture radar (SAR) and laser. In this paper, we use deep convolutional neural network (CNN) denoiser prior based alternating direction method of multipliers (ADMM) algorithm to deblur the image with multiplicative Gamma noise in the plug and play framework. In order to restore the image with more details, a sharpening operator is used. Experimental results show that our method achieves higher evaluation index and more abundant visuals.