{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,18]],"date-time":"2026-11-18T18:59:36Z","timestamp":1795028376154,"version":"4.10.2"},"reference-count":73,"publisher":"Society for Industrial & Applied Mathematics (SIAM)","issue":"4","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["SIAM J. Imaging Sci."],"published-print":{"date-parts":[[2014,1]]},"abstract":"<jats:p>Algorithms for solving variational regularization of ill-posed inverse problems usually involve operators that depend on a collection of continuous parameters. When the operators enjoy some (local) regularity, these parameters can be selected using the so-called Stein Unbiased Risk Estimator (SURE). While this selection is usually performed by an exhaustive search, we address in this work the problem of using the SURE to efficiently optimize for a collection of continuous parameters of the model. When considering nonsmooth regularizers, such as the popular $\\ell_1$-norm corresponding to soft-thresholding mapping, the SURE is a discontinuous function of the parameters preventing the use of gradient descent optimization techniques. Instead, we focus on an approximation of the SURE based on finite differences as proposed by Ramani and Unser for the Monte-Carlo SURE approach. Under mild assumptions on the estimation mapping, we show that this approximation is a weakly differentiable function of the parameters and its weak gradient, coined the Stein Unbiased GrAdient estimator of the Risk (SUGAR), provides an asymptotically (with respect to the data dimension) unbiased estimate of the gradient of the risk. Moreover, in the particular case of soft-thresholding, it is proved to also be a consistent estimator. This gradient estimate can then be used as a basis for performing a quasi-Newton optimization. The computation of the SUGAR relies on the closed-form (weak) differentiation of the nonsmooth function. We provide its expression for a large class of iterative methods including proximal splitting methods and apply our strategy to regularizations involving nonsmooth convex structured penalties. Illustrations of various image restoration and matrix completion problems are given.<\/jats:p>","DOI":"10.1137\/140968045","type":"journal-article","created":{"date-parts":[[2014,11,25]],"date-time":"2014-11-25T12:36:43Z","timestamp":1416919003000},"page":"2448-2487","source":"Crossref","is-referenced-by-count":108,"title":["Stein Unbiased GrAdient estimator of the Risk (SUGAR) for Multiple Parameter Selection"],"prefix":"10.1137","volume":"7","author":[{"given":"Charles-Alban","family":"Deledalle","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Samuel","family":"Vaiter","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jalal","family":"Fadili","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gabriel","family":"Peyr\u00e9","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"351","published-online":{"date-parts":[[2014,11,25]]},"reference":[{"key":"atypb1","doi-asserted-by":"publisher","DOI":"10.1137\/100784114"},{"key":"atypb2","doi-asserted-by":"publisher","DOI":"10.1145\/1944345.1944349"},{"key":"atypb3","doi-asserted-by":"crossref","unstructured":"H. Bauschke and P. L. Combettes,\n                      Convex Analysis and Monotone Operator Theory in Hilbert Spaces\n                      , CMS Books Math.\/Ouvrages Math. SMC, Springer, New York, 2011.","DOI":"10.1007\/978-1-4419-9467-7"},{"key":"atypb4","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2005.857247"},{"key":"atypb5","first-page":"3","author":"Bennett J.","year":"2007","journal-title":"New York"},{"key":"atypb6","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2007.906002"},{"key":"atypb7","doi-asserted-by":"publisher","DOI":"10.1214\/07-AOS538"},{"key":"atypb8","doi-asserted-by":"publisher","DOI":"10.1007\/s10208-009-9045-5"},{"key":"atypb9","doi-asserted-by":"publisher","DOI":"10.1109\/TSP.2013.2270464"},{"key":"atypb10","doi-asserted-by":"publisher","DOI":"10.1007\/s10851-010-0251-1"},{"key":"atypb11","doi-asserted-by":"publisher","DOI":"10.1109\/TSP.2008.921757"},{"key":"atypb12","doi-asserted-by":"publisher","DOI":"10.1109\/JSTSP.2007.910264"},{"key":"atypb13","first-page":"185","author":"Combettes P. L.","year":"2011","journal-title":"New York"},{"key":"atypb14","doi-asserted-by":"publisher","DOI":"10.1137\/050626090"},{"key":"atypb15","doi-asserted-by":"publisher","DOI":"10.1002\/cpa.3160410705"},{"key":"atypb16","doi-asserted-by":"publisher","DOI":"10.1007\/s10851-011-0294-y"},{"key":"atypb17","unstructured":"C.A. Deledalle, G. Peyre\u0301, and J. Fadili,\n                      Stein COnsistent Risk Estimator (SCORE) for Hard Thresholding\n                      , preprint, http:\/\/arxiv.org\/abs\/1301.5874arXiv:1301.5874v1 [math.ST], 2013."},{"key":"atypb18","doi-asserted-by":"crossref","unstructured":"C.A. Deledalle, F. Tupin, and L. Denis,\n                      Poisson NL means: Unsupervised non local means for Poisson noise\n                      , in Proceedings of the 2010 17th IEEE International Conference on Image Processing, 2010, pp. 801-804.","DOI":"10.1109\/ICIP.2010.5653394"},{"key":"atypb19","doi-asserted-by":"publisher","DOI":"10.1088\/1742-6596\/386\/1\/012003"},{"key":"atypb20","unstructured":"C.A. Deledalle, S. Vaiter, G. Peyre\u0301, J. Fadili, and C. Dossal,\n                      Risk Estimation for Matrix Recovery with Spectral Regularization\n                      , preprint, http:\/\/arxiv.org\/abs\/1205.1482arXiv:1205.1482v3 [math.OC], 2012."},{"key":"atypb21","doi-asserted-by":"crossref","unstructured":"C.A. Deledalle, S. Vaiter, G. Peyre\u0301, J. Fadili, and C. Dossal,\n                      Unbiased risk estimation for sparse analysis regularization\n                      , in Proceedings of the 2012 19th IEEE International Conference on Image Processing, 2012, pp. 3053-3056.","DOI":"10.1109\/ICIP.2012.6467544"},{"key":"atypb22","doi-asserted-by":"publisher","DOI":"10.1016\/0370-2693(94)90440-5"},{"key":"atypb23","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.1995.10476626"},{"key":"atypb24","first-page":"809","volume":"23","author":"Dossal C.","year":"2013","journal-title":"Statist. Sinica"},{"key":"atypb25","doi-asserted-by":"publisher","DOI":"10.1137\/100790902"},{"key":"atypb26","unstructured":"A. Edelman,\n                      Jacobians of Matrix Transforms (with Wedge Products\n                      ), Handout 3, MIT, Cambridge, MA, 2005."},{"key":"atypb27","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.1986.10478291"},{"key":"atypb28","doi-asserted-by":"publisher","DOI":"10.1109\/TSP.2008.2008212"},{"key":"atypb29","unstructured":"L. C. Evans and R. F. Gariepy,\n                      Measure Theory and Fine Properties of Functions\n                      , CRC Press, Boca Raton, FL, 1992."},{"key":"atypb30","unstructured":"M. Fazel,\n                      Matrix Rank Minimization with Applications\n                      , Ph.D. thesis, Department of Electrical Engineering, Stanford University, Stanford, CA, 2002."},{"key":"atypb31","unstructured":"D. Gilbarg and N. S. Trudinger,\n                      Elliptic Partial Differential Equations of Second Order\n                      , 2nd ed., Classics Math. 517, Springer-Verlag, Berlin, 1998."},{"key":"atypb32","first-page":"1","volume":"56","author":"Girard D. A.","year":"1989","journal-title":"Math."},{"key":"atypb33","doi-asserted-by":"publisher","DOI":"10.1016\/j.acha.2010.11.005"},{"key":"atypb34","doi-asserted-by":"publisher","DOI":"10.1080\/00401706.1979.10489751"},{"key":"atypb35","doi-asserted-by":"crossref","unstructured":"A. Griewank and A. Walther,\n                      Evaluating Derivatives: Principles and Techniques of Algorithmic Differentiation\n                      , SIAM, Philadelphia, 2008.","DOI":"10.1137\/1.9780898717761"},{"key":"atypb36","unstructured":"J.B. Hiriart-Urruty and C. Lemare\u0301chal,\n                      Convex Analysis and Minimization Algorithms\n                      , Vols. I-II, Springer-Verlag, Berlin, 2001."},{"key":"atypb37","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1176344194"},{"key":"atypb38","doi-asserted-by":"publisher","DOI":"10.1016\/S0167-9473(97)00041-8"},{"key":"atypb39","doi-asserted-by":"publisher","DOI":"10.1080\/03610918908812806"},{"key":"atypb40","doi-asserted-by":"publisher","DOI":"10.1016\/j.jmva.2008.12.002"},{"key":"atypb41","unstructured":"N. Komodakis and J.C. Pesquet,\n                      Playing with Duality: An Overview of Recent Primal-Dual Approaches for Solving Large-Scale Optimization Problems\n                      , preprint, http:\/\/arxiv.org\/abs\/1406.5429arXiv:1406.5429v1 [cs.NA], 2014."},{"key":"atypb42","doi-asserted-by":"publisher","DOI":"10.1007\/s10107-012-0514-2"},{"key":"atypb43","doi-asserted-by":"publisher","DOI":"10.1137\/S089547980036838X"},{"key":"atypb44","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1176349742"},{"key":"atypb45","doi-asserted-by":"publisher","DOI":"10.1109\/TCSVT.2010.2045819"},{"key":"atypb46","unstructured":"S. Mallat,\n                      A Wavelet Tour of Signal Processing. The Sparse Way\n                      , 3rd ed., Elsevier\/Academic Press, Amsterdam, 2009."},{"key":"atypb47","doi-asserted-by":"publisher","DOI":"10.1007\/s10107-006-0042-z"},{"key":"atypb48","doi-asserted-by":"publisher","DOI":"10.1109\/TSP.2009.2026077"},{"key":"atypb49","first-page":"249","author":"Pesquet J.-C.","year":"1997","journal-title":"UK"},{"key":"atypb50","doi-asserted-by":"publisher","DOI":"10.1137\/120872802"},{"key":"atypb51","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2008.2001404"},{"key":"atypb52","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2012.2195015"},{"key":"atypb53","doi-asserted-by":"crossref","unstructured":"S. Ramani, J. Rosen, Z. Liu, and J. A. Fessler,\n                      Iterative weighted risk estimation for nonlinear image restoration with analysis priors\n                      , in Proceedings of SPIE, Vol. 8296, Computational Imaging X, SPIE, Bellingham, WA, 2012, 82960N.","DOI":"10.1117\/12.915643"},{"key":"atypb54","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2008.925392"},{"key":"atypb55","doi-asserted-by":"publisher","DOI":"10.1162\/NECO_a_00076"},{"key":"atypb56","doi-asserted-by":"publisher","DOI":"10.1090\/conm\/059\/10"},{"key":"atypb57","doi-asserted-by":"publisher","DOI":"10.1137\/0314056"},{"key":"atypb58","unstructured":"F. Roosta-Khorasani and U. Ascher,\n                      Improved Bounds on Sample Size for Implicit Matrix Trace Estimators\n                      , preprint, http:\/\/arxiv.org\/abs\/1308.2475arXiv:1308.2475 [cs.NA], 2013."},{"key":"atypb59","doi-asserted-by":"publisher","DOI":"10.1016\/0167-2789(92)90242-F"},{"key":"atypb60","doi-asserted-by":"publisher","DOI":"10.1198\/016214502753479356"},{"key":"atypb61","first-page":"3754","author":"Solo V.","year":"2010","journal-title":"Speech and Signal Processing"},{"key":"atypb62","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1176345632"},{"key":"atypb63","unstructured":"D. Sun and J. Sun,\n                      Nonsmooth Matrix Valued Functions Defined by Singular Values\n                      , Technical report, Department of Decision Sciences, National University of Singapore, Singapore, 2003."},{"key":"atypb64","doi-asserted-by":"publisher","DOI":"10.1214\/11-AOS878"},{"key":"atypb65","doi-asserted-by":"publisher","DOI":"10.1214\/12-AOS1003"},{"key":"atypb66","doi-asserted-by":"publisher","DOI":"10.1016\/j.acha.2012.11.006"},{"key":"atypb67","unstructured":"S. Vaiter, C.A. Deledalle, G. Peyre\u0301, J. M Fadili, and C. Dossal,\n                      The Degrees of Freedom of Partly Smooth Regularizers\n                      , preprint, arXiv:1404.5557v2 [math.ST], 2014."},{"key":"atypb68","doi-asserted-by":"publisher","DOI":"10.1109\/LSP.2009.2027669"},{"key":"atypb69","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2011.2121083"},{"key":"atypb70","doi-asserted-by":"crossref","unstructured":"C. Vonesch, S. Ramani, and M. Unser,\n                      Recursive risk estimation for non-linear image deconvolution with a wavelet-domain sparsity constraint\n                      , in Proceedings of the 15th IEEE International Conference on Image Processing, 2008, pp. 665-668.","DOI":"10.1109\/ICIP.2008.4711842"},{"key":"atypb71","doi-asserted-by":"publisher","DOI":"10.1080\/01621459.1998.10474094"},{"key":"atypb72","doi-asserted-by":"publisher","DOI":"10.1111\/j.1467-9868.2005.00532.x"},{"key":"atypb73","doi-asserted-by":"publisher","DOI":"10.1214\/009053607000000127"}],"container-title":["SIAM Journal on Imaging Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/epubs.siam.org\/doi\/pdf\/10.1137\/140968045","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,21]],"date-time":"2026-08-21T13:14:48Z","timestamp":1787318088000},"score":1,"resource":{"primary":{"URL":"https:\/\/epubs.siam.org\/doi\/10.1137\/140968045"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2014,1]]},"references-count":73,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2014,1]]}},"alternative-id":["10.1137\/140968045"],"URL":"https:\/\/doi.org\/10.1137\/140968045","relation":{},"ISSN":["1936-4954"],"issn-type":[{"value":"1936-4954","type":"electronic"}],"subject":[],"published":{"date-parts":[[2014,1]]}}}