Non-local sparse models for image restoration

Mairal, Julien; Bach, Francis; Ponce, Jean; Sapiro, Guillermo; Zisserman, Andrew

doi:10.1109/iccv.2009.5459452

Cited by 1,429 publications

(1,139 citation statements)

References 27 publications

Supporting

Mentioning

1,129

Contrasting

Unclassified

Order By: Relevance

“…Specifically, in traditional approaches to boundary detection, filters have been applied using linear filtering to generate dense, redundant, representations. Other work has demonstrated that sparse coding can be beneficial for image compression (e.g., Fischer et al, 2006Fischer et al, , 2007Murray and Kreutz-Delgado, 2006;Pece and Petkov, 2000), image restoration (e.g., Elad and Aharon, 2006;Hyvarinen et al, 1998;Mairal et al, 2009), and classification (e.g., Bociu and Pitas, 2004;Mairal et al, 2008;Ramirez et al, 2010;Wright et al, 2009). Here, it has been shown that sparse coding can also be used for boundary detection.…”

Section: Discussionmentioning

confidence: 95%

Image Segmentation Using a Sparse Coding Model of Cortical Area V1

Spratling

2013

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Algorithms that encode images using a sparse set of basis functions have previously been shown to explain aspects of the physiology of primary visual cortex (V1), and have been used for applications such as image compression, restoration, and classification. Here, a sparse coding algorithm, that has previously been used to account of the response properties of orientation tuned cells in primary visual cortex, is applied to the task of perceptually salient boundary detection. The proposed algorithm is currently limited to using only intensity information at a single scale. However, it is shown to out-perform the current state-of-the-art image segmentation method (Pb) when this method is also restricted to using the same information.

show abstract

Section: Discussionmentioning

confidence: 95%

Image Segmentation Using a Sparse Coding Model of Cortical Area V1

Spratling

2013

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

show abstract

“…The two terms in (9) are widely employed in optimization-based dictionary learning (see for example [33]- [38], and the references therein). The first term in (9) imposes an 2 fit between the model and observed data {y it }, and the second term imposes regularization on the dictionary elements {d k } k=1,K , which constitute the columns of D. For the special case in which Ω = I m , the second term in (9) reduces…”

Section: Relationship To Previous Modelsmentioning

confidence: 99%

“…The sparsity manifested via (11) is the most distinctive aspect of the proposed model, relative to previous optimization-based approaches [33]- [38]. In that work one often places shrinkage priors on the weights c i , via 1 regularization γ s i c i 1 ; in such an approach all the terms in (11) are essentially just replaced with γ s i c i 1 .…”

Section: Relationship To Previous Modelsmentioning

confidence: 99%

See 1 more Smart Citation

Coded Hyperspectral Imaging and Blind Compressive Sensing

Rajwade¹,

Kittle²,

Tsai³

et al. 2013

SIAM J. Imaging Sci.

View full text Add to dashboard Cite

Blind compressive sensing (CS) is considered for reconstruction of hyperspectral data imaged by a coded aperture camera. The measurements are manifested as a superposition of the coded wavelengthdependent data, with the ambient three-dimensional hyperspectral datacube mapped to a two-dimensional measurement. The hyperspectral datacube is recovered using a Bayesian implementation of blind CS.Several demonstration experiments are presented, including measurements performed using a coded aperture snapshot spectral imager (CASSI) camera. The proposed approach is capable of efficiently reconstructing large hyperspectral datacubes. Comparisons are made between the proposed algorithm and other techniques employed in compressive sensing, dictionary learning and matrix factorization. Index Termshyperspectral images, image reconstruction, projective transformation, dictionary learning, non-parametric Bayesian, Beta-Bernoulli model, coded aperture snapshot spectral imager (CASSI).

show abstract

“…Given the excellent performance of non-local methods [2,5], learned sparse models [8,18], and the combination of both [6,17] for random Gaussian noise, they were explored for impulse noise as well [20,22]. Non-local methods use redundant visual information within an image (i.e., self-similarity) to group similar image patches together, followed by collaborative filtering [2,5].…”

Section: Related Workmentioning

confidence: 99%