Article ID Journal Published Year Pages File Type
529713 Journal of Visual Communication and Image Representation 2016 10 Pages PDF
Abstract

•A new simple primal–dual method is proposed.•The method can get a closed-form solution in each iteration.•O(1/N) convergence rate in the ergodic sense is established.•The proposed method outperforms some state-of-the-art optimization algorithms.•The algorithm is easy to extended to work on more general total variation image problems.

In this study we propose a simple primal–dual method for total variation minimization problems. A predictor–corrector scheme to the dual variable is used in our algorithms and convergence of the method is proved. We also show that the iterative scheme has O(1/N)O(1/N) convergence rate in the ergodic sense, where N denotes the iteration number. Numerical results including image deblurring and computerized tomography reconstruction demonstrate the efficient of the new algorithms.

Related Topics
Physical Sciences and Engineering Computer Science Computer Vision and Pattern Recognition
Authors
, , ,