Edge-preserving and scale-dependent properties of total variation regularization

By David M. Strong, Tony F. Chan. 2003

673 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/8605e0131d086ee65dd862dd2046df06696df8c4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Image up-sampling using total-variation regularization with a new observation model (2005, background): “The effect of the values of on total-variation regularization is analyzed in [37].”
Total variation models for variable lighting face recognition (2006, methodology): “Just like many other TV-based models (e.g., the ROF model [31]), the TV-L 1 model keeps the sharp object edges in u [39] because minimizing the TV regularization term penalizes all intensity variations in a uniform way (in contrast to min R jruj 2 , which penalizes larger changes, like sharp edges,…”
Recovering missing pixels for Landsat ETM + SLC-off imagery using multi-temporal regression analysis and a regularization method (2013, background): “Generally, the latter two priors give a better performance in edge-preserving but take a longer time (Strong & Chan, 2003).”
Total Variation Blind Deconvolution: The Devil Is in the Details (2014, methodology): “The proof exploits recent work of Condat [5], Strong and Chan [20] and the taut string algorithm of Davies and Kovacs [6].”
Modern regularization methods for inverse problems (2018, background): “[112, 343, 331, 120, 322, 342, 77, 104, 4]), computational optimization techniques for nonsmooth problems (cf.”
Algorithms for Finding Global Minimizers of Image Segmentation and Denoising Models (2006, background): “This choice can be facilitated by the observation that λ acts as a scale parameter [25]: Its value determines in some sense the smallest image feature that will be maintained in the reconstructed image.”
An Iterative Regularization Method for Total Variation-Based Image Restoration (2005, methodology): “This follows from an analysis of the ROF model by Strong and Chan [42].”
Iterative Weighted Maximum Likelihood Denoising With Probabilistic Patch-Based Weights (2009, background): “Markovian priors are local in essence, and lead to stronger attenuation of several small disconnected regions than that of a single region [4].”
Multifaceted Feature Visualization: Uncovering the Different Types of Features Learned By Each Neuron in Deep Neural Networks (2016, background): “S1 for hypothesis as to why this may occur. preserving the sharpness of edges (Strong et al., 2003).”
NL-SAR: A Unified Nonlocal Framework for Resolution-Preserving (Pol)(In)SAR Denoising (2015, background): “wavelets) or loss of contrast (with total variation minimization) [17].”
Aspects of Total Variation Regularized L[sup 1] Function Approximation (2005, methodology): “The results of [5] generalize the results of [22], but also show that the class of binary images that have this weaker property (i.e. images for which the solution to the standard ROF model turns out to be a constant multiple of the observed image) is still rather limited; for example, there are…”
Recent Developments in Total Variation Image Restoration (2004, background): “This generalization has the beneﬁt that vector valued images deﬁned on the line whose components are monotone functions with identical boundary conditions all have the same energy, regardless of their smoothness.”
A New Alternating Minimization Algorithm for Total Variation Image Reconstruction (2008, background): “The reader interested in this topic is referred to [39].”
A proximal decomposition method for solving convex variational inverse problems (2008, background): “Experiment 2 In image recovery, variational formulations involving total variation [15, 44, 47] or sparsity promoting potentials [5, 8, 14, 28] are popular.”
SimPEG: An open source framework for simulation and gradient based parameter estimation in geophysical applications (2015, background): “The l2 norms referred to above are appropriate for many problems, however models norms, such as lp-norms, total variation, minimum support stabilizing functionals, or rotated smoothness operators that favor different character and/or include additional information can also be designed (cf. Oldenburg”
Face Recognition Under Varying Illumination Using Gradientfaces (2009, methodology): “proposed to use the [31] model to obtain the smoothed in image logarithm domain, which utilized the edge-preserving capability of total variation model.”