Convex Variational Image Restoration with Histogram Priors

@article{Swoboda2013ConvexVI,
  title={Convex Variational Image Restoration with Histogram Priors},
  author={Paul Swoboda and Christoph Schn{\"o}rr},
  journal={SIAM J. Imaging Sci.},
  year={2013},
  volume={6},
  pages={1719-1735}
}
We present a novel variational approach to image restoration (e.g., denoising, inpainting, labeling) that enables us to complement established variational approaches with a histogram-based prior, enforcing closeness of the solution to some given empirical measure. By minimizing a single objective function, the approach utilizes simultaneously two quite different sources of information for restoration: spatial context in terms of some smoothness prior and nonspatial statistics in terms of the… 

Figures from this paper

Wasserstein Loss for Image Synthesis and Restoration
TLDR
A novel variational approach to imposing statistical constraints on the output of both image generation and image restoration (for instance, to achieve denoising and inpainting) methods is presented.
RWRM: Residual Wasserstein regularization model for image restoration
TLDR
A residual Wasserstein regularization model (RWRM) is proposed, in which a residual histogram constraint is subtly embedded into a type of variational minimization problems, utilizing the Wassersteins distance from the optimal transport theory.
Texture Reconstruction Guided by a High-Resolution Patch
TLDR
This paper proposes a variational method that combines two approaches that are texture synthesis and image reconstruction that leads to a significant improvement, both visually and numerically, with respect to the state-of-the-art algorithms for solving similar problems.
Adaptive color transfer with relaxed optimal transport
TLDR
This work extends the work of [1] by modeling the spatial distribution of colors within the image domain and tuning automatically the relaxation parameters and demonstrates the capacity of the model to adapt itself to the considered data.
Imaging with Kantorovich-Rubinstein Discrepancy
TLDR
A variational regularisation model endowed with a Kantorovich-Rubinstein discrepancy term and total variation regularization in the context of image denoising and cartoon-texture decomposition is discussed and connections to several other recently proposed methods are pointed out.
Transport Between RGB Images Motivated by Dynamic Optimal Transport
TLDR
This work proposes two discrete variational models for the interpolation between RGB images, a constrained and a penalized one which can also handle unbalanced transport and shows that a minimizer of this discrete model exists, but it is not unique for some special initial/final images.
A Smoothed Dual Approach for Variational Wasserstein Problems
TLDR
The dual formulation of Wasserstein variational problems introduced recently can be regularized using an entropic smoothing, which leads to smooth, differentiable, convex optimization problems that are simpler to implement and numerically more stable.
Dynamic optimal transport with mixed boundary condition for color image processing
TLDR
This paper shows how the transportation problem for RGB color images can be tackled by prescribing periodic boundary conditions in the color dimension and sketches how the same idea can be used in a modified way to transport periodic 1D data such as the histogram of cyclic hue components of images.
Optimal Transport for Image Processing
Optimal Transport is a well developed mathematical theory that defines robust metrics between probability distributions. The computation of optimal displacements between densities through the
Transport Optimal pour le Traitement d'Images
Optimal Transport is a well developed mathematical theory that defines robust metrics between probability distributions. The computation of optimal displacements between densities through the
...
...

References

SHOWING 1-10 OF 33 REFERENCES
Global Solutions of Variational Models with Convex Regularization
TLDR
An algorithmic framework for computing global solutions of variational models with convex regularity terms that permit quite arbitrary data terms and a fast primal-dual algorithm which significantly outperforms existing algorithms is proposed.
Wasserstein regularization of imaging problem
TLDR
A novel and generic framework embedding statistical constraints for variational problems by combining the proposed approximate Wasserstein constraint on color statistics with a generic geometric-based regularization term in a unified variational minimization problem.
An Experimental Comparison of Discrete and Continuous Shape Optimization Methods
TLDR
A quantitative comparison of the reconstruction accuracy and computation times is presented which allows to assess some of the strengths and limitations of both approaches and a novel method to approximate length regularity in a graph cut based framework.
A First-Order Primal-Dual Algorithm for Convex Problems with Applications to Imaging
  • A. ChambolleT. Pock
  • Mathematics, Computer Science
    Journal of Mathematical Imaging and Vision
  • 2010
TLDR
A first-order primal-dual algorithm for non-smooth convex optimization problems with known saddle-point structure can achieve O(1/N2) convergence on problems, where the primal or the dual objective is uniformly convex, and it can show linear convergence, i.e. O(ωN) for some ω∈(0,1), on smooth problems.
Continuous Multiclass Labeling Approaches and Algorithms
TLDR
This work proposes a globally convergent Douglas-Rachford scheme and shows that a sequence of dual iterates can be recovered in order to provide a posteriori optimality bounds and shows competitive performance on synthetic and real-world images.
Histogram Based Segmentation Using Wasserstein Distances
TLDR
A new nonparametric region-based active contour model for clutter image segmentation is proposed and to quantify the similarity between two clutter regions, it is proposed to compare their respective histograms using the Wasserstein distance.
Nonlinear total variation based noise removal algorithms
Discrete and Continuous Models for Partitioning Problems
TLDR
A method is proposed to significantly reduce costs in a fully automatic way for a large class of metrics including tree metrics, thus generalizing a method recently proposed by Strekalovskiy and Cremers.
Static and Dynamic Texture Mixing Using Optimal Transport
TLDR
This paper tackles the problem of mixing static and dynamic texture by combining the statistical properties of an input set of images or videos by defining the distance between texture models, derive the geodesic path, and define the barycenter between several texture models.
Nonlocal Operators with Applications to Image Processing
TLDR
This topic can be viewed as an extension of spectral graph theory and the diffusion geometry framework to functional analysis and PDE-like evolutions to define new types of flows and functionals for image processing and elsewhere.
...
...