Image Completion by Diffusion Maps and Spectral Relaxation

  title={Image Completion by Diffusion Maps and Spectral Relaxation},
  author={Shai Gepshtein and Yosi Keller},
  journal={IEEE Transactions on Image Processing},
We present a framework for image inpainting that utilizes the diffusion framework approach to spectral dimensionality reduction. We show that on formulating the inpainting problem in the embedding domain, the domain to be inpainted is smoother in general, particularly for the textured images. Thus, the textured images can be inpainted through simple exemplar-based and variational methods. We discuss the properties of the induced smoothness and relate it to the underlying assumptions used in… 

Exemplar-based image inpainting using svd-based approximation matrix and multi-scale analysis

A new inpainting technique to overcome the limitation of the lack of global analysis on the image by using an approximation matrix, which is in fact a gray-scale copy of the original image in which the target region is approximated throughout the process of rank lowering.

Gradient-based low rank method and its application in image inpainting

A new image prior is proposed, which implies the low rank prior knowledge of image gradients, which works well for both structural and texture images and outperforms other techniques, in terms of both objective and subjective performance measures.


An image inpainting scheme has been proposed that utilizes the spatial contextual information approach for image completion and the method’s efficacy is embodied by real images.

Image Inpainting Through Quality Based Patch Selection Matrix

A modified Criminisi’s exempler based image-inpainting method is proposed in which quality based patch selection is performed in the process of inpainting of digital images to recover the lost part of the image in a visually plausible way such that the changes made to the image are not detected by the normal user.

Image Inpainting Based on Image Mapping and Object Removal Using Semi-Automatic Method

  • BaZasaheb H. PatilP. Patil
  • Computer Science
    2018 International Conference On Advances in Communication and Computing Technology (ICACCT)
  • 2018
The main importance of the approach is to utilize association of a pixel with its neighbors to get precise smoothness over the inpainted image.

Fast Exemplar-Based Image Inpainting Using a New Pruning Technique

A new algorithm for fast exemplar-based image inpainting by using image pyramid and a novel pruning scheme is presented, which is much more efficient computationally and shown to compare favorably with contemporary state-of-the-art schemes.

Study of Image Recovery by Object and Noise Removal by Using Enhanced Inpainting Techniques

In this paper a method for image denoising and inpainting is proposed and the benefit of this approach is that it is easier to inpaint.

Manifold Learning via the Principle Bundle Approach

An image denoising algorithm is introduced, called the diffusive vector non-local Euclidean median (dvNLEM), by combining the traditional nonlocal Euclidesan median, the rotational structure in the patch space, and the diffusion distance.

Structure-aware image inpainting using patch scale optimization

Photo Realistic Image Completion via Dense Correspondence

A hierarchical PatchMatch method that progressively estimates a dense correspondence, which is able to capture small deformations between images, which can achieve photo realistic image completion results.



A Comprehensive Framework for Image Inpainting

This paper combines copy-and-paste texture synthesis, geometric partial differential equations (PDEs), and coherence among neighboring pixels in a variational model, and provides a working algorithm for image inpainting trying to approximate the minimum of the proposed energy functional.

Vector-valued image regularization with PDEs: a common framework for different applications

A unifying expression is proposed that gathers the majority of PDE-based formalisms for vector-valued image regularization into a single generic anisotropic diffusion equation, allowing us to implement the authors' regularization framework with accuracy by taking the local filtering properties of the proposed equations into account.

Simultaneous structure and texture image inpainting

The novel contribution of the paper is the combination of these three previously developed components: image decomposition with inpainting and texture synthesis, which permits the simultaneous use of filling-in algorithms that are suited for different image characteristics.

Region filling and object removal by exemplar-based image inpainting

The simultaneous propagation of texture and structure information is achieved by a single, efficient algorithm that combines the advantages of two approaches: exemplar-based texture synthesis and block-based sampling process.

Strong-continuation, contrast-invariant inpainting with a third-order optimal PDE

  • M. Bertalmío
  • Computer Science
    IEEE Transactions on Image Processing
  • 2006
A novel third-order PDE is derived, which, in both its accuracy and contrast invariance, outperforms the approaches cited above.

Image inpainting

A novel algorithm for digital inpainting of still images that attempts to replicate the basic techniques used by professional restorators, and does not require the user to specify where the novel information comes from.

Nonlinear Poisson Image Completion using Color Manifold

A novel approach, termed as nonlinear Poisson completion, to solve the problem of completing large-scale missing parts on complex background with nonlinearly progressive color changes by embedding nonlinear subspace of progressive colors in color manifold.

Euler's Elastica and Curvature-Based Inpainting

A computational scheme based on numerical PDEs is presented, which allows the automatic handling of topologically complex inpainting domains and connects to the earlier works of Bertalmio, Sapiro, Caselles, and Ballester.

Sparse Representation for Color Image Restoration

This work puts forward ways for handling nonhomogeneous noise and missing information, paving the way to state-of-the-art results in applications such as color image denoising, demosaicing, and inpainting, as demonstrated in this paper.