A nonparametric minimum entropy image deblurring algorithm

In this paper we address the image restoration problem in the variational framework. Classical approaches minimize the L^p norm of the residual and rely on parametric assumptions on the noise statistical model. We relax this parametric hypothesis and we formulate the problem on the basis of nonparametric density estimates. The proposed approach minimizes the residual differential entropy. Experimental results with non gaussian distributions show the interest of such a nonparametric approach. Images quality is evaluated by means of the PSNR measure and SSIM index, more adapted to the human visual system.