Image Restoration via Simultaneous Sparse Coding and Gaussian Scale Mixture
Weisheng Dong,Guangming Shi
2014-01-01
Abstract:Sparse coding has been shown to be relevant to both variational and Bayesian approaches. The regularization parameter in variational image restoration is intrinsically connected with the shape parameter of sparse coefficients’ distribution in Bayesian methods. How to tune those parameters in a principled yet spatially adaptive fashion turns out to be a challenging problem especially for the class of nonlocal image models. In this work, we propose an empirical Bayesian approach of addressing this issue more specifically, a nonlocal extension of Gaussian scale mixture (GSM) model is developed based on simultaneous sparse coding (SSC) and its applications into image restoration are explored. It is shown that the variances of sparse coefficients (the field of scalar multipliers of Gaussians) if treated as a latent variable can be jointly estimated along with the unknown sparse coefficients via the method of alternating optimization. Unlike previous attacks to Bayesian image restoration, ours leads to closed-form solutions involving iterative shrinkage/filtering only and therefore admits computationally efficient implementation. Our solution can be viewed as an exemplar of demonstrating a new variational approach W. Dong School of Electronic Engineering, Xidian University, Xi’an, 710071, China E-mail: wsdong@mail.xidian.edu.cn G. Shi School of Electronic Engineering, Xidian University, Xi’an, 710071, China E-mail: gmshi@xidian.edu.cn Yi Ma Visual Computing Group, Microsoft Research of Asia, Beijing, China E-mail: mayi@microsoft.com Xin Li Lane Department of CSEE, West Virginia University, Morgantown, WV 26506-6109 USA E-mail: xin.li@ieee.org toward empirical Bayesian inference with parametric models, which has computational advantages over the existing variational Bayesian approaches. When applied to image restoration, our experimental results have shown that the proposed SSC-GSM technique can both preserve the sharpness of edges and suppress undesirable artifacts. Thanks to its capability of achieving a better spatial adaptation, SSC-GSM based image restoration often delivers reconstructed images with higher subjective/objective qualities than other competing approaches.