We propose a new image denoising algorithm by incorporating the $L_1$-regularization technique and the least squares method. In order to improve the quality of the reconstructed images, we adopt a high order least squares method along with new iterative nonlocal weights. In particular, we devise a measurement to estimate the nonlocal similarities between patches by using both data values and their derivatives. Some experimental results are presented to demonstrate the capability of the proposed algorithm.

This presentation is part of Contributed Presentation “CP7 - Contributed session 7”

Authors:

Byeongseon Jeong (Institute of Mathematical Sciences, Ewha Womans University, Seoul 120-750)

Yunjin Park (Department of Mathematics, Ewha Womans University, Seoul 120-750)

Jungho Yoon (Department of Mathematics, Ewha Womans University, Seoul 120-750)

Hyoseon Yang (Institute of Mathematical Sciences, Ewha Womans University)

Keywords:

$l_1$ minimization, computer vision, image denoising, image reconstruction, nonlinear optimization, nonlocal similarity