×

FRIST-flipping and rotation invariant sparsifying transform learning and applications. (English) Zbl 1380.94039

Summary: Features based on sparse representation, especially using the synthesis dictionary model, have been heavily exploited in signal processing and computer vision. However, synthesis dictionary learning typically involves NP-hard sparse coding and expensive learning steps. Recently, sparsifying transform learning received interest for its cheap computation and its optimal updates in the alternating algorithms. In this work, we develop a methodology for learning flipping and rotation invariant sparsifying transforms, dubbed FRIST, to better represent natural images that contain textures with various geometrical directions. The proposed alternating FRIST learning algorithm involves efficient optimal updates. We provide a convergence guarantee, and demonstrate the empirical convergence behavior of the proposed FRIST learning approach. Preliminary experiments show the promising performance of FRIST learning for sparse image representation, segmentation, denoising, robust inpainting, and compressed sensing-based magnetic resonance image reconstruction.

MSC:

94A08 Image processing (compression, reconstruction, etc.) in information and communication theory
65K10 Numerical optimization and variational techniques
68T05 Learning and adaptive systems in artificial intelligence
90C90 Applications of mathematical programming
PDFBibTeX XMLCite
Full Text: DOI arXiv

References:

[1] Michael Elad’s Homepage online: www.cs.technion.ac.il/ elad/Various/KSVD_Matlab_ToolBox.zip (Accessed: 6 November 2015)
[2] The USC-SIPI Image Database online: http://sipi.usc.edu/database/database.php?volume=misc; (Accessed: 6 November 2015)
[3] Transform Learning Website online: http://transformlearning.csl.illinois.edu/ (Accessed: 1 February 2016)
[4] Aharon M and Elad M 2008 Sparse and redundant modeling of image content using an image-signature-dictionary SIAM J. Imaging Sci.1 228-47 · Zbl 1182.68337 · doi:10.1137/07070156X
[5] Aharon M, Elad M and Bruckstein A 2006 K-SVD: an algorithm for designing overcomplete dictionaries for sparse representation IEEE Trans. Signal Process.54 4311-22 · Zbl 1375.94040 · doi:10.1109/TSP.2006.881199
[6] Bruckstein A M, Donoho D L and Elad M 2009 From sparse solutions of systems of equations to sparse modeling of signals and images SIAM Rev.51 34-81 · Zbl 1178.68619 · doi:10.1137/060657704
[7] Bryt O and Elad M 2008 Compression of facial images using the k-svd algorithm J. Vis. Commun. Image Represent.19 270-82 · doi:10.1016/j.jvcir.2008.03.001
[8] Chen S S, Donoho D L and Saunders M A 1998 Atomic decomposition by basis pursuit SIAM J. Sci. Comput.20 33-61 · Zbl 0919.94002 · doi:10.1137/S1064827596304010
[9] Dabov K, Foi A, Katkovnik V and Egiazarian K 2007 Image denoising by sparse 3D transform-domain collaborative filtering IEEE Trans. Image Process.16 2080-95 · doi:10.1109/TIP.2007.901238
[10] Davis G, Mallat S and Avellaneda M 1997 Adaptive greedy approximations J. Constructive Approx.13 57-98 · Zbl 0885.41006 · doi:10.1007/BF02678430
[11] Elad M and Aharon M 2006 Image denoising via sparse and redundant representations over learned dictionaries IEEE Trans. Image Process.15 3736-45 · doi:10.1109/TIP.2006.881969
[12] Elad M, Milanfar P and Rubinstein R 2007 Analysis versus synthesis in signal priors Inverse Problems23 947-68 · Zbl 1138.93055 · doi:10.1088/0266-5611/23/3/007
[13] Engan K, Aase S and Hakon-Husoy J 1999 Method of optimal directions for frame design Proc. IEEE Int. Conf. on Acoustics, Speech, and Signal Processing pp 2443-6
[14] Gleichman S and Eldar Y C 2011 Blind compressed sensing IEEE Trans. Inf. Theory57 6958-75 · Zbl 1365.94095 · doi:10.1109/TIT.2011.2165821
[15] Ke Y and Sukthankar R 2004 PCA-SIFT: a more distinctive representation for local image descriptors IEEE Computer Society Conf. on Computer Vision and Pattern Recognitionvol 2 pp 506-13
[16] Lowe D G 1999 Object recognition from local scale-invariant features IEEE Int. Conf. on Computer visionvol 2 pp 1150-7
[17] Lustig M, Donoho D and Pauly J M 2007 Sparse MRI: the application of compressed sensing for rapid MR imaging Magn. Reson. Med.58 1182-95 · doi:10.1002/mrm.21391
[18] Mairal J, Elad M and Sapiro G 2008 Sparse representation for color image restoration IEEE Trans. Image Process.17 53-69 · doi:10.1109/TIP.2007.911828
[19] Mairal J, Bach F, Ponce J and Sapiro G 2010 Online learning for matrix factorization and sparse coding J. Mach. Learn. Res.11 19-60 · Zbl 1242.62087
[20] Mallat S 1999 A Wavelet Tour of Signal Processing (London: Academic)
[21] Mallat S G and Zhang Z 1993 Matching pursuits with time-frequency dictionaries IEEE Trans. Signal Process.41 3397-415 · Zbl 0842.94004 · doi:10.1109/78.258082
[22] Ning B, Qu X, Guo D, Hu C and Chen Z 2013 Magnetic resonance image reconstruction using trained geometric directions in 2D redundant wavelets domain and non-convex optimization Magn. Reson. Imaging31 1611-22 · doi:10.1016/j.mri.2013.07.010
[23] Pati Y, Rezaiifar R and Krishnaprasad P 1993 Orthogonal matching pursuit : recursive function approximation with applications to wavelet decomposition Asilomar Conf. on Signals, Systems and Computervol 1 pp 40-4
[24] Pennec E L and Mallat S 2005 Bandelet image approximation and compression Multiscale Modeling Simul.4 992-1039 · Zbl 1091.94006 · doi:10.1137/040619454
[25] Pfister L and Bresler Y 2014 Adaptive sparsifying transforms for iterative tomographic reconstruction Int. Conf. on Image Formation in X-Ray Computed Tomography
[26] Pfister L and Bresler Y 2015 Learning sparsifying filter banks Proc. SPIE9597 9597-2 · doi:10.1117/12.2188663
[27] Pratt W K, Kane J and Andrews H C 1969 Hadamard transform image coding Proc. IEEE57 58-68
[28] Qu X, Guo D, Ning B, Hou Y, Lin Y, Cai S and Chen Z 2012 Undersampled MRI reconstruction with patch-based directional wavelets Magn. Reson. Imaging30 964-77 · doi:10.1016/j.mri.2012.02.019
[29] Qu X, Hou Y, Lam F, Guo D, Zhong J and Chen Z 2014 Magnetic resonance image reconstruction from undersampled measurements using a patch-based nonlocal operator Med. Image Anal.18 843-56 · doi:10.1016/j.media.2013.09.007
[30] Ram I, Elad M and Cohen I 2013 Image processing using smooth ordering of its patches IEEE Trans. Image Process.22 2764-74 · Zbl 1373.94339 · doi:10.1109/TIP.2013.2257813
[31] Ravishankar S and Bresler Y 2013 Learning doubly sparse transforms for images IEEE Trans. Image Process.22 4598-612 · Zbl 1373.94344 · doi:10.1109/TIP.2013.2274384
[32] Ravishankar S and Bresler Y 2013 Learning overcomplete sparsifying transforms for signal processing IEEE Int. Conf. on Acoustics, Speech and Signal Processing pp 3088-92
[33] Ravishankar S and Bresler Y 2013 Learning sparsifying transforms IEEE Trans. Signal Process.61 1072-86 · Zbl 1393.94409 · doi:10.1109/TSP.2012.2226449
[34] Ravishankar S and Bresler Y 2015 Efficient blind compressed sensing using sparsifying transforms with convergence guarantees and application to magnetic resonance imaging SIAM J. Imaging Sci.8 2519-57 · Zbl 1343.94013 · doi:10.1137/141002293
[35] Ravishankar S, Wen B and Bresler Y 2015 Online sparsifying transform learning—part I: algorithms IEEE J. Sel. Top. Signal Process.9 625-36 · doi:10.1109/JSTSP.2015.2417131
[36] Ravishankar S and Bresler Y 2011 MR image reconstruction from highly undersampled k-space data by dictionary learning IEEE Trans. Med. Imaging30 1028-41 · doi:10.1109/TMI.2010.2090538
[37] Ravishankar S and Bresler Y 2014 ℓ0 sparsifying transform learning with efficient optimal updates and convergence guarantees IEEE Trans. Signal Process.63 2389-404 · Zbl 1394.94477 · doi:10.1109/TSP.2015.2405503
[38] Rubinstein R, Bruckstein A M and Elad M 2010 Dictionaries for sparse representation modeling Proc. IEEE98 1045-57 · doi:10.1109/JPROC.2010.2040551
[39] Skretting K and Engan K 2010 Recursive least squares dictionary learning algorithm IEEE Trans. Signal Process.58 2121-30 · Zbl 1392.68365 · doi:10.1109/TSP.2010.2040671
[40] Trzasko J and Manduca A 2009 Highly undersampled magnetic resonance image reconstruction via homotopic-minimization IEEE Trans. Med. Imaging28 106-21 · doi:10.1109/TMI.2008.927346
[41] Watson D 2013 Contouring: a Guide to the Analysis and Display of Spatial Data (Amsterdam: Elsevier)
[42] Wen B, Li Y and Bresler Y 2017 When sparsity meets low-rankness: transform learning with non-local low-rank constraint for image restoration IEEE Int. Conf. on Acoustics, Speech and Signal Processing · doi:10.1109/ICASSP.2017.7952566
[43] Wen B, Ravishankar S and Bresler Y 2014 Learning overcomplete sparsifying transforms with block cosparsity IEEE Int. Conf. on Image Processing pp 803-7
[44] Wen B, Ravishankar S and Bresler Y 2015 Structured overcomplete sparsifying transform learning with convergence guarantees and applications Int. J. Comput. Vis.114 137-67 · Zbl 1398.94053 · doi:10.1007/s11263-014-0761-1
[45] Wersing H, Eggert J and Körner E 2003 Artificial Neural Networks and Neural Information Processing pp 385-92
[46] Yaghoobi M, Blumensath T and Davies M 2009 Dictionary learning for sparse approximations with the majorization method IEEE Trans. Signal Process.57 2178-91 · Zbl 1391.94456 · doi:10.1109/TSP.2009.2016257
[47] Yang T 1986 Finite Element Structural Analysis vol 2 (Englewood Cliffs, NJ: Prentice Hall)
[48] Yu G and Sapiro G 2011 DCT image denoising: a simple and effective image denoising algorithm Image Process. Line1 292-6 · doi:10.5201/ipol.2011.ys-dct
[49] Yu G, Sapiro G and Mallat S 2010 Image modeling and enhancement via structured sparse model selection IEEE Int. Conf. on Image Processing pp 1641-4
[50] Yu G, Sapiro G and Mallat S 2012 Solving inverse problems with piecewise linear estimators: from gaussian mixture models to structured sparsity IEEE Trans. Image Process.21 2481-99 · Zbl 1373.94471 · doi:10.1109/TIP.2011.2176743
[51] Zelnik-Manor L, Rosenblum K and Eldar Y 2012 Dictionary optimization for block-sparse representations IEEE Trans. Signal Process.60 2386-95 · Zbl 1393.94731 · doi:10.1109/TSP.2012.2187642
[52] Zhan Z, Cai J, Guo D, Liu Y, Chen Z and Qu X 2016 Fast multi-class dictionaries learning with geometrical directions in MRI reconstruction IEEE Trans. Biomed. Eng.63 1850-61 · doi:10.1109/TBME.2015.2503756
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.