Adaptive total variation model for image denoising based on modified orientation information measure

In this article, an adaptive total variation model by selecting the most appropriate generalized coefficient p adaptively based on modified orientation information measure is introduced. The model can keep the balance between noise smoothing and edges preserving adaptively. In the past, the solutions of TV model were based on nonlinear partial differential equations (PDEs) and the resulting algorithms were very complicated. Therefore we present a promoted effective algorithm based on Bregman iterative regularization for solving the adaptive TV minimization problems in image denoising with no involving solving PDEs. Experimental results show that the proposed denoising model and effective algorithm can properly preserve the main information of the original image with fast solving convergence rate, while the PSNR and subjective visual effect of the denoising images are improved significantly.