Patch ordering as a regularization for inverse problems in image processing
From MaRDI portal
Abstract: Recent work in image processing suggests that operating on (overlapping) patches in an image may lead to state-of-the-art results. This has been demonstrated for a variety of problems including denoising, inpainting, deblurring, and super-resolution. The work reported in [1,2] takes an extra step forward by showing that ordering these patches to form an approximate shortest path can be leveraged for better processing. The core idea is to apply a simple filter on the resulting 1D smoothed signal obtained after the patch-permutation. This idea has been also explored in combination with a wavelet pyramid, leading eventually to a sophisticated and highly effective regularizer for inverse problems in imaging. In this work we further study the patch-permutation concept, and harness it to propose a new simple yet effective regularization for image restoration problems. Our approach builds on the classic Maximum A'posteriori probability (MAP), with a penalty function consisting of a regular log-likelihood term and a novel permutation-based regularization term. Using a plain 1D Laplacian, the proposed regularization forces robust smoothness (L1) on the permuted pixels. Since the permutation originates from patch-ordering, we propose to accumulate the smoothness terms over all the patches' pixels. Furthermore, we take into account the found distances between adjacent patches in the ordering, by weighting the Laplacian outcome. We demonstrate the proposed scheme on a diverse set of problems: (i) severe Poisson image denoising, (ii) Gaussian image denoising, (iii) image deblurring, and (iv) single image super-resolution. In all these cases, we use recent methods that handle these problems as initialization to our scheme. This is followed by an L-BFGS optimization of the above-described penalty function, leading to state-of-the-art results, and especially so for highly ill-posed cases.
Recommendations
- Non-local regularization of inverse problems
- The Little Engine that Could: Regularization by Denoising (RED)
- A unified view on patch aggregation
- Adaptive morphologic regularizations for inverse problems
- The lower bound of nonlocal gradient for non-convex and non-smooth image patches based regularization
Cites work
- A General Framework for Regularized, Similarity-Based Image Restoration
- An Effective Heuristic Algorithm for the Traveling-Salesman Problem
- BM3D Frames and Variational Image Deblurring
- Boosting of image denoising algorithms
- Generalized Tree-Based Wavelet Transform
- Image Processing Using Smooth Ordering of its Patches
- Image Super-Resolution Via Sparse Representation
- Introduction to algorithms
- Local and nonlocal discrete regularization on weighted graphs for image and mesh processing
- Non-local regularization of inverse problems
- Nonlocal discrete p-Laplacian driven image and manifold processing
- Nonlocally Centralized Sparse Representation for Image Restoration
- Numerical Optimization
- On the limited memory BFGS method for large scale optimization
- Patch-Ordering-Based Wavelet Frame and Its Use in Inverse Problems
- Poisson noise reduction with non-local PCA
- Progressive Image Denoising Through Hybrid Graph Laplacian Regularization: A Unified Framework
- Solving Inverse Problems With Piecewise Linear Estimators: From Gaussian Mixture Models to Structured Sparsity
- Sparsity-Based Poisson Denoising With Dictionary Learning
- Symmetric Smoothing Filters From Global Consistency Constraints
Cited in
(5)- Registration and inpainting of biological slice images based on Bendlet transform and interval Shannon-cosine wavelet
- Equivalence of weak and strong modes of measures on topological vector spaces
- Image Denoising: The Deep Learning Revolution and Beyond—A Survey Paper
- A unified view on patch aggregation
- Smoothing randomized block-coordinate proximal gradient algorithms for nonsmooth nonconvex composite optimization
This page was built for publication: Patch ordering as a regularization for inverse problems in image processing
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2797782)
