Research Article  Open Access
Deblurring by Solving a TV^{p}Regularized Optimization Problem Using Split Bregman Method
Abstract
Image deblurring is formulated as an unconstrained minimization problem, and its penalty function is the sum of the error term and TV^{p}regularizers with . Although TV^{p}regularizer is a powerful tool that can significantly promote the sparseness of image gradients, it is neither convex nor smooth, thus making the presented optimization problem more difficult to deal with. To solve this minimization problem efficiently, such problem is first reformulated as an equivalent constrained minimization problem by introducing new variables and new constraints. Thereafter, the split Bregman method, as a solver, splits the new constrained minimization problem into subproblems. For each subproblem, the corresponding efficient method is applied to ensure the existence of closedform solutions. In simulated experiments, the proposed algorithm and some stateoftheart algorithms are applied to restore three types of blurrednoisy images. The restored results show that the proposed algorithm is valid for image deblurring and is found to outperform other algorithms in experiments.
1. Introduction
Over the past half century, image deblurring has been intensively studied and extensively applied in many fields, such as biometric identification, remote sensing, and video surveillance, among others. As an illposed and inverse problem, image deblurring requires a stable solution. For this purpose, Tikhonov et al. [1] propose regularization technology that initially uses as the regularizer, with as the Tikhonov operator. However, the Tikhonov regularized image deblurring algorithms favor reconstructing oversmoothing images such that the most important edges are lost. To preserve edges, Rudin et al. [2] propose the total variation (TV) model: , where TV() is a TVregularizer that has two forms, namely, isotropic TVregularizer () and anisotropic TVregularizer (). From the piecewise continuity of natural images, the values of the elements in the gradients of images are mostly equal to zero; that is, the gradients of natural images are sparse. As a powerful tool, the TVregularizer has been employed by many stateoftheart image deblurring algorithms [3–5], but recent research reveals that for modeling the sparseness of image gradient, the norm () with is more suitable than the norm () of TVregularizer [6]. Moreover, the regularizer outperforms the TVregularizer in terms of edge preservation and noise suppression.
From the nonconvexity and nonsmoothness of norm, imaging inverse problems involving regularizer or regularizer are generally confronted with challenges in the existence of solutions and low efficiencies of corresponding algorithms. For regularized image deblurring with specific values of , Levin et al. [7] originally use iterative reweighted least squares (IRLS) to decompose the deblurring problem into equivalent subproblems, for which the conjugate gradient (CG) is then employed. Zhuo et al. [8] then propose a similar algorithm for blur identification in blind deblurring, which also adopts IRLS method with CG iteration. When applying IRLS to image deblurring problems involving regularizers, hundreds of CG iterations are usually needed, thus giving rise to negative effects on the speed of image deblurring. For regularized compressive sensing, a commonly used method is interior point, which can solve the corresponding optimization problems in polynomial time under certain conditions [9]. However, the interiorpoint method is inapplicable to optimization problems with equality constraints, and the initial points are difficult to choose. These factors can be regarded as the two primary inherent drawbacks of the interiorpoint method.
Variable splitting technology has recently been proposed and extensively applied to inverse problems [10, 11], including regularized image deblurring problems [12]. By introducing new variables under the framework of variable splitting, the optimization tasks of imaging inverse problems are transformed into subtasks that are easier to address. For each subtask, the corresponding solver should be carefully selected to ensure that the additional computational costs are low.
Although regularizer is more appropriate than TVregularizer in promoting the sparseness of image gradients, statistics show that very few image deblurring algorithms have considered the regularizer. The most likely reason is that when regularizers are involved, image deblurring problems are neither convex nor smooth, and efficient algorithms are difficult to develop. To take full advantage of regularizer, this study models image deblurring as an unconstrained minimization problem, the penalty function of which is combined with the error term and the sparse regularizers with . For convenience of image deblurring, the unconstrained minimization problem is reexpressed as an equivalent constrained minimization problem with new variables and constraints. Thereafter, the constrained minimization problem is decomposed into subproblems by the split Bregman method [13]. For these subproblems, generalized shrinkage/thresholding (GST) functions [14], fast Fourier transforms (FFTs), and other methods are introduced. By restoring different types of blurrednoisy images, the proposed algorithm shows effectiveness and better performance than some similar stateoftheart algorithms (Algorithm 1). The proposed algorithm is empirically found to converge to satisfactory solutions with only several iterations, as verified by experiments.
2. Family of Bregman Methods
As variants of the standard Bregman iteration method, the split Bregman method (SBM), linearized Bregman method (LBM) [15], and Bregmanized operator splitting (BOS) [3] are the key members of the family of Bregman methods. SBM is proposed for the following minimization problem: where is generally nonsmooth, and is a linear operator that may be invertible or not. Involving new variables and constraints, problem (1) is reformulated in the following equivalent form: SBM is employed to solve problem (2) by splitting it into the following subproblems: where the constant . For subproblems (3) to (5), if is a convex function, closedform solutions generally exist. In particular, when and , the solutions to subproblem (4) are the popular softthresholding function and hardthresholding function, respectively. When solving some other types of optimization problems, SBM may also be called the alternating direction method of multipliers (ADMM) or DouglasRachford splitting method. More details on the connections among SBM, ADMM, and the DouglasRachford splitting method can be found in the literature [16].
When applied to optimization problems, such as TV optimization problem or optimization problem, SBM is generally faster than the twostep iterative thresholding methods [17] and Nesterov’s methods [18, 19]. Therefore, the applications of SBM to imaging inverse problems have become active research fields. Through SBM, many classic image deblurring/denoising problems [13, 20, 21] have been efficiently solved. Aside from SBM, LBM and BOS are also popular methods for imaging inverse problems. LBM, which can be regarded as the combination of the standard Bregman iteration and fixedpoint methods, is initially proposed for compressed sensing problems, and its application is later extended to image deblurring. To solve problem (1), with , LBM generates the following subproblems: where the distance is defined as
If in iteration, “kicking” [22] is used in LBM. However, LBM cannot be applied when problem (1) involves TVregularizer or multiple regularizers. BOS is proposed for TV and nonlocal TV (NLTV) models in imaging inverse problems. BOS decomposes the problem (1) into the following subproblems: The forwardbackward operator splitting (FBOS) [23] is then used to solve problem (8) as follows: BOS can be regarded as Bregman iteration plus FBOS because (8) and (9) represent the standard Bregman iteration. Compared with SBM, which can usually obtain good results with only one inner iteration, LBM and BOS are more dependent on the inner solvers or inner iterations. Therefore, when applied to image deblurring problems, both LBM and BOS show lower efficiencies than SBM. In addition to the fast convergent rate, the parameters of SBM need not vary with iteration. Thus, the number of conditions can be reduced by carefully selecting the parameters. In particular, SBM is more suitable for multiregularized image deblurring problems.
3. The Proposed Algorithm: SBMTV^{p}
The degradation of images can be modeled as follows: where , , , and represent blurrednoisy image, blur operator, the unknown sharp image in vector form, and Gaussian additive noise, respectively. cannot obviously be estimated according only to (11), especially when is heavily illposed. Therefore, as a rule, image deblurring problems are transformed into optimization problems. This study casts image deblurring as follows: where and denote the vertical and horizontal gradient operators, respectively; is regularizer with ; and the constant controls the tradeoff between data fidelity and regularization. By introducing the variables and and the constraints and , problem (12) can be reformulated as follows:
From the nonsmoothness of norm, problem (13) cannot be analytically solved in a direct manner. If we let , , , , and in problem (2), then we can find that problems (2) and (13) are the same problems with the same structures. Therefore, this study uses SBM as an “indirect” method of solving problem (13). By SBM, problem (13) is split into the following subproblems: where the constant .
Except for subproblems (15) and (16), other subproblems can be directly computed, and subproblem (14) (i.e., subproblem) has a closedform solution: where and . According to the theory of Tikhonov regularization, (19) provides a regularized solution that is an approximation of the exact solution . Choosing the proper values for , with , we find that is approaching the exact solution [1]. With the help of FFTs in (19), , , , and can all be computed with the computational cost of . After precomputing , , , and , and with the equal help of FFTs, can be obtained with the cost of . Several numerical methods, such as IRLS and interiorpoint method, have been applied to the minimization of subproblems (15) and (16) with specific values of , but these methods cannot usually converge to satisfactory solutions, especially for largescale problems (e.g., image deblurring). To guarantee the convergence of minimization of subproblems (15) and (16), as well as to reduce the costs of computing them, GST functions are employed as the solvers. When we let to approach and let and represent any elements of and , respectively, then and can be analytically computed by
The GST function in (20) is defined as where is the signum function; ; is iteratively computed by and the thresholding value . When , the GST function is which is the famous softthresholding function. When , the GST function is given by which is the famous hardthresholding function. Therefore, softthresholding and hardthresholding functions can be regarded as the special cases of the GST function. The computational costs of both functions are because and are computed elementbyelement.
The proposed iterative algorithm is generated from (17) to (22). The proposed algorithm is called “” because it solves regularized image deblurring with the use of the split Bregman method. The efficiency of the proposed algorithm is mainly dependent on the computations of subproblems , , and , and all have low computational costs as briefly analyzed above. Therefore, the efficiency of the proposed algorithm is high, as will be verified by the experiments. As mentioned in [14], the GST function has the following four properties: , if , , and for . Thus, according to the proof in [24], the convergence of GST is guaranteed; that is, subproblems and converge with iteration.
4. Experiments and Results
As shown in Figure 1, four standard gray images of different types are chosen as the sharp images for experiments. To create the blurrednoisy images shown in Figure 2, according the degradation model in (11), sharp images are firstly filtered by blur operators in Table 1, where fspecial is MATLAB function, and then Gaussian noises of different levels are added to these images. In Figure 2, blurred signaltonoise ratio (BSNR) is defined as where is the variance of noise. BSNR is employed as the objective criterion of degradation. The proposed is employed to restore the blurrednoisy images in Figure 2 to verify its effectiveness. The algorithms in [3, 7] are introduced for comparison in terms of restoring the same blurrednoisy images to verify the superiority of . The experiments are conducted on a laptop with Windows XP, Intel Duo 2 CPU @ 2.10 GHz, and 2 GB of RAM, and all the algorithms are implemented in MATLAB R2011b platform. For , the values of , , and are set as 2, , and , respectively. With only several iterations, can obtain satisfactory results. Thus, is set to be 10. As for the algorithms in [3, 7], all parameters are left at the default settings. The stopping criteria of DeblurSps and _BOS are “ and ” and “ or ( is the noise level),” respectively, which are the default criteria adopted by their references. and denote outer iterations and inner iterations (CG iterations), respectively. The restored results of all algorithms are shown in Figures 3 to 5 and Tables 3, 4, and 5, with peak signaltonoise ratio (PSNR) defined as PSNR is employed as the objective criterion for the quality of restored images. For , several typical values, that is, , , and , are considered. “DeblurSps” represents the algorithm in [7], which solves the regularized () image deblurring problem using IRLS and CG. “TV_BOS” and “NLTV_BOS” represent the algorithms in [3], which apply the Bregmanized operator splitting to the TV and NLTV models. To compare the speeds of all algorithms, each of them is run for 10 times to restore the blurrednoisy images, and the average time of each algorithm is shown in Table 2.





(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m)
(n)
(o)
(p)
(q)
(r)
(s)
(t)
(u)
(v)
(w)
(x)
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m)
(n)
(o)
(p)
(q)
(r)
(s)
(t)
(u)
(v)
(w)
(x)
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m)
(n)
(o)
(p)
(q)
(r)
(s)
(t)
(u)
(v)
(w)
(x)
The results in Figures 3 to 5 and Tables 3 to 5 demonstrate that can efficiently restore different types of blurrednoisy images and displays superior performance over other algorithms, especially in terms of speed. As an approximation method, IRLS cannot ensure that DeblurSps will obtain the optimal solution to the image deblurring problem. In addition, to solve a sequence of subproblems generated by IRLS, hundreds of CG iterations are needed. Therefore, the restored results of DeblurSps are worse than those of . Moreover, this algorithm is slower than and TV_BOS. However, as a regularized algorithm, DeblurSps outperforms TV_BOS and NLTV_BOS on the restored results. When BOS is applied to TVregularized and NLTVregularized image deblurring problems, the resulting subproblems cannot be solved by simple methods (e.g., shrinkage methods), such that the process of image deblurring becomes more complex with the loss of efficiency. In particular, for the NLTVregularized image deblurring problem, additional time is required to update the weight function. Therefore, as shown in Tables 2 to 5 and Figures 3 to 5, the algorithms in [3] are inferior to other algorithms in terms of their performances on the restored results and their speed. The convergence of cannot be theoretically proven. Therefore, to analyze its convergence, Figure 6 shows the evolution of in the procedure of iteration, where denotes the Frobenius norm. The curves clearly demonstrate that with iteration, is approaching , that is, the converges. In general, with two to four iterations, can obtain satisfactory solutions.
(a)
(b)
(c)
5. Conclusions
This study proposes a novel image deblurring algorithm that solves a regularized minimization problem using SBM. To obtain the closedform solutions to the subproblems generated by SBM efficiently, the FFTs and GST functions are introduced. In the experiment, three types of blurred noisy images are restored by and several stateoftheart algorithms. Comparison results show that outperforms other algorithms in terms of restored results and speed. Although this study does not prove the convergence of , Figure 6 clearly illustrates the fact.
Conflict of Interests
The author declares that there is no conflict of interests regarding the publication of this paper.
References
 A. N. Tikhonov, A. Goncharsky, V. V. Stepanov, and A. G. Yagola, Numerical Methods for the Solution of IllPosed Problems, Kluwer Academic, Dordrecht, The Netherlands, 1995. View at: MathSciNet
 L. I. Rudin, S. Osher, and E. Fatemi, “Nonlinear total variation based noise removal algorithms,” Physica D: Nonlinear Phenomena, vol. 60, no. 1–4, pp. 259–268, 1992. View at: Publisher Site  Google Scholar
 X. Zhang, M. Burger, X. Bresson, and S. Osher, “Bregmanized nonlocal regularization for deconvolution and sparse reconstruction,” SIAM Journal on Imaging Sciences, vol. 3, no. 3, pp. 253–276, 2010. View at: Publisher Site  Google Scholar  MathSciNet
 Y.W. Wen and R. H. Chan, “Parameter selection for totalvariationbased image restoration using discrepancy principle,” IEEE Transactions on Image Processing, vol. 21, no. 4, pp. 1770–1781, 2012. View at: Publisher Site  Google Scholar  MathSciNet
 T. S. Cho, C. L. Zitnick, N. Joshi, S. B. Kang, R. Szeliski, and W. T. Freeman, “Image restoration by matching gradient distributions,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 34, no. 4, pp. 683–694, 2012. View at: Publisher Site  Google Scholar
 M. Nikolova, M. K. Ng, S. Zhang, and W.K. Ching, “Efficient reconstruction of piecewise constant images using nonsmooth nonconvex minimization,” SIAM Journal on Imaging Sciences, vol. 1, no. 1, pp. 2–25, 2008. View at: Publisher Site  Google Scholar  MathSciNet
 A. Levin, R. Fergus, F. Durand, and W. T. Freeman, “Image and depth from a conventional camera with a coded aperture,” ACM Transactions on Graphics, vol. 26, no. 3, Article ID 1276464, 2007. View at: Publisher Site  Google Scholar
 S. Zhuo, D. Guo, and T. Sim, “Robust flash deblurring,” in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR '10), pp. 2440–2447, San Francisco, Calif, USA, June 2010. View at: Publisher Site  Google Scholar
 D. Ge, X. Jiang, and Y. Ye, “A note on the complexity of ${L}_{p}$ minimization,” Mathematical Programming, vol. 129, no. 2, pp. 285–299, 2011. View at: Publisher Site  Google Scholar  MathSciNet
 A. Danielyan, V. Katkovnik, and K. Egiazarian, “BM3D frames and variational image deblurring,” IEEE Transactions on Image Processing, vol. 21, no. 4, pp. 1715–1728, 2012. View at: Publisher Site  Google Scholar  MathSciNet
 D. Goldfarb and S. Ma, “Fast multiplesplitting algorithms for convex optimization,” SIAM Journal on Optimization, vol. 22, no. 2, pp. 533–556, 2012. View at: Publisher Site  Google Scholar  MathSciNet
 L. Xu, S. Zheng, and J. Jia, “Unnatural L_{0} sparse representation for natural image deblurring,” in Proceedings of the 26th IEEE Conference on Computer Vision and Pattern Recognition (CVPR '13), pp. 1107–1114, Portland, Ore, USA, June 2013. View at: Publisher Site  Google Scholar
 T. Goldstein and S. Osher, “The split Bregman method for L_{1}regularized problems,” SIAM Journal on Imaging Sciences, vol. 2, no. 2, pp. 323–343, 2009. View at: Publisher Site  Google Scholar  MathSciNet
 W. Zuo, D. Meng, L. Zhang, X. Feng, and D. Zhang, “A generalized iterated shrinkage algorithm for nonconvex sparse coding,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV '13), pp. 217–224, Sydney, Australia, December 2013. View at: Publisher Site  Google Scholar
 W. Yin, “Analysis and generalizations of the linearized Bregman model,” SIAM Journal on Imaging Sciences, vol. 3, no. 4, pp. 856–877, 2010. View at: Publisher Site  Google Scholar  MathSciNet
 S. Setzer, “Operator splittings, Bregman methods and frame shrinkage in image processing,” International Journal of Computer Vision, vol. 92, no. 3, pp. 265–280, 2011. View at: Publisher Site  Google Scholar  MathSciNet
 A. Beck and M. Teboulle, “Fast gradientbased algorithms for constrained total variation image denoising and deblurring problems,” IEEE Transactions on Image Processing, vol. 18, no. 11, pp. 2419–2434, 2009. View at: Publisher Site  Google Scholar  MathSciNet
 C. Chaux, J.C. Pesquet, and N. Pustelnik, “Nested iterative algorithms for convex constrained image recovery problems,” SIAM Journal on Imaging Sciences, vol. 2, no. 2, pp. 730–762, 2009. View at: Publisher Site  Google Scholar  MathSciNet
 S. Becker, J. Bobin, and E. J. Candès, “NESTA: a fast and accurate firstorder method for sparse recovery,” SIAM Journal on Imaging Sciences, vol. 4, no. 1, pp. 1–39, 2011. View at: Publisher Site  Google Scholar  MathSciNet
 J. Cai, H. Ji, C. Liu, and Z. Shen, “Blind motion deblurring from a single image using sparse approximation,” in Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition, pp. 104–111, Miami, Fla, USA, 2009. View at: Google Scholar
 S. Setzer, G. Steidl, and T. Teuber, “Deblurring Poissonian images by split Bregman techniques,” Journal of Visual Communication and Image Representation, vol. 21, no. 3, pp. 193–199, 2010. View at: Publisher Site  Google Scholar
 S. Osher, Y. Mao, B. Dong, and W. Yin, “Fast linearized Bregman iteration for compressive sensing and sparse denoising,” Communications in Mathematical Sciences, vol. 8, no. 1, pp. 93–111, 2010. View at: Publisher Site  Google Scholar  MathSciNet
 E. T. Hale, W. Yin, and Y. Zhang, “Fixedpoint continuation for ${l}_{1}$minimization: methodology and convergence,” SIAM Journal on Optimization, vol. 19, no. 3, pp. 1107–1130, 2008. View at: Publisher Site  Google Scholar  MathSciNet
 Y. She, “Thresholdingbased iterative selection procedures for model selection and shrinkage,” Electronic Journal of Statistics, vol. 3, pp. 384–415, 2009. View at: Publisher Site  Google Scholar  MathSciNet
Copyright
Copyright © 2014 Su Xiao. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.