idamnjanovic@1:
idamnjanovic@29: SMALLbox Version 1.0 beta
idamnjanovic@1:
idamnjanovic@29: 11th May, 2010
idamnjanovic@29:
idamnjanovic@29: Version 1.0 beta of SMALLbox is the release candidate that is distributed
idamnjanovic@29: for testing and bug fixing purposes. Please send all bugs, requests and suggestions to:
idamnjanovic@29:
idamnjanovic@29: ivan.damnjanovic@elec.qmul.ac.uk
idamnjanovic@1:
idamnjanovic@1:
idamnjanovic@1: ---------------------------------------------------------------------------
idamnjanovic@1:
idamnjanovic@29: Copyright (2010): Ivan Damnjanovic, Matthew Davies
idamnjanovic@29: Centre for Digital Music,
idamnjanovic@29: Queen Mary University of London
idamnjanovic@1:
idamnjanovic@1: SMALLbox is distributed under the terms of the GNU General Public License 3
idamnjanovic@1:
idamnjanovic@1: This program is free software: you can redistribute it and/or modify
idamnjanovic@1: it under the terms of the GNU General Public License as published by
idamnjanovic@1: the Free Software Foundation, either version 3 of the License, or
idamnjanovic@1: (at your option) any later version.
idamnjanovic@1:
idamnjanovic@1: This program is distributed in the hope that it will be useful,
idamnjanovic@1: but WITHOUT ANY WARRANTY; without even the implied warranty of
idamnjanovic@1: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
idamnjanovic@1: GNU General Public License for more details.
idamnjanovic@1:
idamnjanovic@1: You should have received a copy of the GNU General Public License
idamnjanovic@1: along with this program. If not, see <http://www.gnu.org/licenses/>.
idamnjanovic@1:
idamnjanovic@1: ---------------------------------------------------------------------------
idamnjanovic@1:
idamnjanovic@1: SMALLbox is a MATLAB Evaluation Framework for the EU FET-OPEN Project, no: 225913:
idamnjanovic@1: Sparse Models, Algorithms, and Learning for Large-scale data (SMALL).
idamnjanovic@1:
idamnjanovic@1: The project team includes researchers from the following institutions:
idamnjanovic@1:
idamnjanovic@1: INRIA: http://www.irisa.fr/metiss/gribonval/
idamnjanovic@1: University of Edinburgh: http://www.see.ed.ac.uk/~mdavies4/
idamnjanovic@1: Queen Mary, University of London: http://www.elec.qmul.ac.uk/people/markp/
idamnjanovic@1: EPFL: http://people.epfl.ch/pierre.vandergheynst
idamnjanovic@1: Technion: http://www.cs.technion.ac.il/~elad/
idamnjanovic@1:
idamnjanovic@1: ---------------------------------------------------------------------------
idamnjanovic@1:
idamnjanovic@29: will download and install the following existing toolboxes
idamnjanovic@1: related to Sparse Representations, Compressed Sensing and Dictionary Learning:
idamnjanovic@1:
idamnjanovic@1:
idamnjanovic@29: - SPARCO (v.1.2) - set of sparse representation problems[5]
idamnjanovic@29: - SparseLab (v.2.1) - set of sparse solvers [1]
idamnjanovic@29: - Sparsify (v.0.4) - set of greedy and hard thresholding algorithms [2]
idamnjanovic@29: - SPGL1 (v.1.7) - large-scale sparse reconstruction solver [3]
idamnjanovic@29: - GPSR (v.6.0) - Gradient projection for sparse reconstruction [4]
idamnjanovic@29: - KSVD-box (v.13) and OMP-box (v.10) - dictionary learning [6]
idamnjanovic@29: - KSVDS-box (v.11) and OMPS-box (v.1) - sparse dictionary learning [7].
idamnjanovic@1:
idamnjanovic@1:
idamnjanovic@1:
idamnjanovic@1: IMPORTANT:
idamnjanovic@1: In order to use SparseLab please register at
idamnjanovic@1: http://cgi.stanford.edu/group/sparselab/cgi-bin/register.pl
idamnjanovic@1:
idamnjanovic@1: IMPORTANT:
idamnjanovic@1: To successfully install all toolboxes you will need to have MEX setup to compile C files.
idamnjanovic@1: If this is not already setup, run "mex -setup" or type "help mex" in the MATLAB command prompt.
idamnjanovic@1:
idamnjanovic@1: IMPORTANT:
idamnjanovic@1: Because the toolboxes are downloaded automatically, you must have an internet connection
idamnjanovic@1: to successfully install SMALLbox.
idamnjanovic@29:
idamnjanovic@29: IMPORTANT:
idamnjanovic@29: If you are running Matlab on MAC OSX or Linux, you must start Matlab with the jvm enabled.
idamnjanovic@1: Not doing so, will prevent you being able to unzip the downloaded toolboxes.
idamnjanovic@1:
idamnjanovic@1: To install the toolbox run the command "SMALLboxsetup" from the MATLAB command prompt.
idamnjanovic@1:
idamnjanovic@29: Once installed, there are two optional demo functions that can be run:
idamnjanovic@29:
idamnjanovic@29: Example test of solvers from different toolboxes on Sparco compressed sensing problems and
idamnjanovic@29: Example test of dictionary learning techniques on image denoising problems
idamnjanovic@29:
idamnjanovic@29: Further examples can be found in {SMALLbox root}/examples directory
idamnjanovic@1:
idamnjanovic@1: ---------------------------------------------------------------------------
idamnjanovic@1:
idamnjanovic@1: For more information on the SMALL Project, please visit the following website:
idamnjanovic@1:
idamnjanovic@29: http://small-project.eu
idamnjanovic@1:
idamnjanovic@1:
idamnjanovic@1: Contacts: ivan.damnjanovic@elec.qmul.ac.uk
idamnjanovic@1: matthew.davies@elec.qmul.ac.uk
idamnjanovic@1:
idamnjanovic@1:
idamnjanovic@1: This code is in experimental stage; any comments or bug reports are
idamnjanovic@29: very welcome. More information about using SMALLbox will be includied in release version
idamnjanovic@29: documenation file.
idamnjanovic@29:
idamnjanovic@29: References:
idamnjanovic@29:
idamnjanovic@29: 1. Donoho, D., Stodden, V., Tsaig, Y.: Sparselab. 2007, http://sparselab.stanford.edu/
idamnjanovic@29: 2. Blumensath, T., Davies, M. E.: Gradient pursuits.
idamnjanovic@29: In IEEE Transactions on Signal Processing, vol. 56, no. 6, pp. 2370�2382, June 2008.
idamnjanovic@29: 3. Berg, E. v., Friedlander, M. P.: Probing the pareto frontier for basis pursuit solutions.
idamnjanovic@29: In SIAM Journal on Scientific Computing, vol. 31, no. 2, pp. 890�912, 2008.
idamnjanovic@29: 4. Figueiredo, M. A. T., Nowak, R. D., Wright, S. J.: Gradient projection for sparse
idamnjanovic@29: reconstruction: Application to compressed sensing and other inverse problems. In Journal
idamnjanovic@29: of Selected Topics in Signal Processing:Special Issue on Convex Optimization for Signal Processing, December 2007.
idamnjanovic@29: 5. Berg, E. v., Friedlander, M. P., Hennenfent, G., Herrmann, F., Saab, R., Yilmaz, O.: Sparco:
idamnjanovic@29: A testing framework for sparse reconstruction. In ACM Trans. on Mathematical Software, 35(4):1-16, February 2009.
idamnjanovic@29: 6. Aharon, M., Elad, M., Bruckstein, A. M.: The K-SVD: An algorithm for designing of overcomplete
idamnjanovic@29: dictionaries for sparse representation. In IEEE Transactions on Signal Processing, vol. 54, no. 11, pp. 4311�4322, 2006.
idamnjanovic@29: 7. Rubinstein, R., Zibulevsky, M. and Elad, M.: Double Sparsity: Learning Sparse Dictionaries
idamnjanovic@29: for Sparse Signal Approximation. In IEEE Transactions on Signal Processing, Vol. 58, No. 3, Pages 1553-1564, March 2010.