Statistics > Machine Learning
[Submitted on 3 Nov 2014 (v1), revised 30 Dec 2014 (this version, v2), latest version 30 Dec 2017 (v3)]
Title:Modular proximal optimization for multidimensional total-variation regularization
View PDFAbstract:One of the most frequently used notions of "structured sparsity" is that of sparse (discrete) gradients, a structure typically elicited through \emph{Total-Variation (TV)} regularizers. This paper focuses on anisotropic TV-regularizers, in particular on $\ell_p$-norm \emph{weighted TV regularizers} for which it develops efficient algorithms to compute the corresponding proximity operators. Our algorithms enable one to scalably incorporate TV regularization of vector, matrix, or tensor data into a proximal convex optimization solvers. For the special case of vectors, we derive and implement a highly efficient weighted 1D-TV solver. This solver provides a backbone for subsequently handling the more complex task of higher-dimensional (two or more) TV by means of a modular proximal optimization approach. We present numerical experiments that demonstrate how our 1D-TV solver matches or exceeds the best known 1D-TV solvers. Thereafter, we illustrate the benefits of our modular design through extensive experiments on: (i) image denoising; (ii) image deconvolution; and (iii) four variants of fused-lasso. Our results show the flexibility and speed our TV solvers offer over competing approaches. To underscore our claims, we provide our TV solvers in an easy to use multi-threaded C++ library (which also aids reproducibility of our results).
Submission history
From: Suvrit Sra [view email][v1] Mon, 3 Nov 2014 18:08:44 UTC (8,499 KB)
[v2] Tue, 30 Dec 2014 11:22:36 UTC (6,903 KB)
[v3] Sat, 30 Dec 2017 20:00:21 UTC (7,282 KB)
Current browse context:
stat.ML
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.