PROJECT TITLE :

Adaptive Image Denoising by Targeted Databases - 2015

ABSTRACT:

We tend to propose a data-dependent denoising procedure to restore noisy pictures. Different from existing denoising algorithms that rummage around for patches from either the noisy image or a generic database, the new algorithm finds patches from a database that contains relevant patches. We formulate the denoising downside as an optimal filter design downside and make 2 contributions. 1st, we tend to confirm the premise function of the denoising filter by solving a group sparsity minimization downside. The optimization formulation generalizes existing denoising algorithms and offers systematic analysis of the performance. Improvement strategies are proposed to reinforce the patch search method. Second, we have a tendency to confirm the spectral coefficients of the denoising filter by considering a localized Bayesian prior. The localized prior leverages the similarity of the targeted database, alleviates the intensive Bayesian computation, and links the new technique to the classical linear minimum mean squared error estimation. We tend to demonstrate applications of the proposed method in a variety of scenarios, together with text images, multiview images, and face pictures. Experimental results show the prevalence of the new algorithm over existing strategies.