Mercurial > hg > smallbox

changeset 129:5a6cef81227e

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Merge from branch "ivand_dev"
author Ivan Damnjanovic lnx <ivan.damnjanovic@eecs.qmul.ac.uk>
date Mon, 13 Jun 2011 14:58:00 +0100
parents 6f78b069e541 (current diff) 8e660fd14774 (diff)
children d8783debe482
files
diffstat 36 files changed, 282 insertions(+), 189 deletions(-) [+]
line wrap: on
line diff
--- a/DL/RLS-DLA/SMALL_rlsdla.m	Wed May 25 15:34:37 2011 +0100
+++ b/DL/RLS-DLA/SMALL_rlsdla.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,5 @@
-function Dictionary = SMALL_rlsdla(X, params)
-%% Recursive Least Squares Dictionary Learning Algorithm
+function [D] = SMALL_rlsdla(X, params)
+%%  Recursive Least Squares Dictionary Learning Algorithm
 %
 %   D = SMALL_rlsdla(X, params) - runs RLS-DLA algorithm for
 %   training signals specified as columns of matrix X with parameters 
@@ -200,6 +200,5 @@
    end
 end
 
-Dictionary = D;
 
 end
--- a/DL/my_dummy_DL.m	Wed May 25 15:34:37 2011 +0100
+++ b/DL/my_dummy_DL.m	Mon Jun 13 14:58:00 2011 +0100
@@ -0,0 +1,22 @@
+function [D]=my_dummy_DL(X, params) 
+%%  Template function that can be used for DL algorithm implementation 
+%   
+%   
+%   input arguments:    
+%                       X - matrix columns are training signals
+%                       params - struycture with parameters for DL
+%                       algorithm
+%
+%   output arguments: 
+%                       D - Learned dictionary.
+
+%%  Change copyright notice as appropriate:
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2009 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.
+%%
--- a/Problems/ImgDenoise_reconstruct.m	Wed May 25 15:34:37 2011 +0100
+++ b/Problems/ImgDenoise_reconstruct.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,10 @@
 function reconstructed=ImgDenoise_reconstruct(y, Problem, SparseDict)
-%%%  Pierre Villars Example - reconstruction function
+%%  Image Denoising Problem reconstruction function
+%   
+%   This reconstruction function is using sparse representation y 
+%   in dictionary Problem.A to reconstruct the patches of the denoised
+%   image.
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -9,12 +14,8 @@
 %   published by the Free Software Foundation; either version 2 of the
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
-%   
-%   This example is based on the experiment suggested by Professor Pierre
-%   Vandergheynst on the SMALL meeting in Villars.
+%%
 
-%   using sparse representation y in dictionary Problem.A reconstruct the
-%   patches from the target image
 
 % stepsize %
 if (isfield(Problem,'stepsize'))
--- a/Problems/Pierre_reconstruct.m	Wed May 25 15:34:37 2011 +0100
+++ b/Problems/Pierre_reconstruct.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,11 @@
 function reconstructed=Pierre_reconstruct(y, Problem)
-%%%  Pierre Villars Example - reconstruction function
+%%  Pierre Villars Example - reconstruction function
+%
+%   This example is based on the experiment suggested by Professor Pierre
+%   Vandergheynst on the SMALL meeting in Villars.
+%   The function is using sparse representation y in dictionary Problem.A 
+%   to reconstruct the patches of the target image.
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -9,12 +15,10 @@
 %   published by the Free Software Foundation; either version 2 of the
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
-%   
-%   This example is based on the experiment suggested by Professor Pierre
-%   Vandergheynst on the SMALL meeting in Villars.
+%%   
 
-%   using sparse representation y in dictionary Problem.A reconstruct the
-%   patches from the target image
+
+
 
 imout=Problem.A*y;
 
--- a/Problems/generateAMT_Learning_Problem.m	Wed May 25 15:34:37 2011 +0100
+++ b/Problems/generateAMT_Learning_Problem.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,16 +1,6 @@
 function data = generateAMT_Learning_Problem(nfft, windowSize, overlap)
-%%% Generate Automatic Music Transcription Problem
+%%  Generate Automatic Music Transcription Problem
 %   
-%   Centre for Digital Music, Queen Mary, University of London.
-%   This file copyright 2009 Ivan Damnjanovic.
-%
-%   This program is free software; you can redistribute it and/or
-%   modify it under the terms of the GNU General Public License as
-%   published by the Free Software Foundation; either version 2 of the
-%   License, or (at your option) any later version.  See the file
-%   COPYING included with this distribution for more information.
-%   
-%
 %   generateAMT_Learning_Problem is a part of the SMALLbox and generates
 %   a problem that can be used for comparison of Dictionary Learning/Sparse
 %   Representation techniques in automatic music transcription scenario.
@@ -31,7 +21,16 @@
 %       notesOriginal - notes of the original audio to be used for
 %                       comparison (if midi of the original exists)
 %       name - name of the audio file to transcribe
-    
+
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2009 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.
+%  
 %%
 FS=filesep;
 if ~ exist( 'nfft', 'var' ) || isempty(nfft), nfft = 4096; end
--- a/Problems/generateAudioDenoiseProblem.m	Wed May 25 15:34:37 2011 +0100
+++ b/Problems/generateAudioDenoiseProblem.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,16 +1,5 @@
 function data=generateAudioDenoiseProblem(au, trainnum, blocksize, dictsize, overlap, sigma, gain, maxval, initdict);
-%%% Audio Denoising Problem
-%
-%   Centre for Digital Music, Queen Mary, University of London.
-%   This file copyright 2010 Ivan Damnjanovic.
-%
-%   This program is free software; you can redistribute it and/or
-%   modify it under the terms of the GNU General Public License as
-%   published by the Free Software Foundation; either version 2 of the
-%   License, or (at your option) any later version.  See the file
-%   COPYING included with this distribution for more information.
-%   
-% 
+%%  Audio Denoising Problem - needs revision, not yet finalised
 %
 %   generateAudioDenoiseProblem is part of the SMALLbox and generate a
 %   problem for comaprison of Dictionary Learning/Sparse Representation
@@ -24,7 +13,15 @@
 %       overlap - ammount of overlaping frames between 0 and 1
 %   
 
-
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2010 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.
+%%  
 disp(' ');
 disp('  **********  Denoising Problem  **********');
 disp(' ');
--- a/Problems/generateImageDenoiseProblem.m	Wed May 25 15:34:37 2011 +0100
+++ b/Problems/generateImageDenoiseProblem.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,5 @@
 function data=generateImageDenoiseProblem(im, trainnum, blocksize, dictsize, sigma, gain, maxval, initdict);
-%% Generate Image Denoising Problem
+%%  Generate Image Denoising Problem
 %   
 %   generateImageDenoiseProblem is a part of the SMALLbox and generates
 %   a problem that can be used for comparison of Dictionary Learning/Sparse
--- a/examples/Automatic Music Transcription/SMALL_AMT_DL_test.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Automatic Music Transcription/SMALL_AMT_DL_test.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,9 @@
-%%  DICTIONARY LEARNING FOR AUTOMATIC MUSIC TRANSCRIPTION EXAMPLE 1
+%%  Dictionary Learning for Automatic Music Transcription - KSVD vs SPAMS
 %   
+%   *WARNING!* You should have SPAMS in your search path in order for this
+%   script to work.Due to licensing issues SPAMS can not be automatically 
+%   provided in SMALLbox (http://www.di.ens.fr/willow/SPAMS/downloads.html).
+%  
 %   This file contains an example of how SMALLbox can be used to test diferent
 %   dictionary learning techniques in Automatic Music Transcription problem.
 %   It calls generateAMT_Learning_Problem that will let you to choose midi,
--- a/examples/Automatic Music Transcription/SMALL_AMT_KSVD_Err_test.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Automatic Music Transcription/SMALL_AMT_KSVD_Err_test.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,14 +1,10 @@
-%% DICTIONARY LEARNING FOR AUTOMATIC MUSIC TRANSCRIPTION
+%%  Dictionary Learning for Automatic Music Transcription - KSVD residual 
+%%  error test
+%   
+%   *WARNING!* You should have SPAMS in your search path in order for this
+%   script to work.Due to licensing issues SPAMS can not be automatically 
+%   provided in SMALLbox (http://www.di.ens.fr/willow/SPAMS/downloads.html).
 %
-%   Centre for Digital Music, Queen Mary, University of London.
-%   This file copyright 2010 Ivan Damnjanovic.
-%
-%   This program is free software; you can redistribute it and/or
-%   modify it under the terms of the GNU General Public License as
-%   published by the Free Software Foundation; either version 2 of the
-%   License, or (at your option) any later version.  See the file
-%   COPYING included with this distribution for more information.
-%   
 %   This file contains an example of how SMALLbox can be used to test diferent
 %   dictionary learning techniques in Automatic Music Transcription problem.
 %   It calls generateAMT_Learning_Problem that will let you to choose midi,
@@ -20,6 +16,15 @@
 %   dictionary learning technique defined in DL structure.
 %
 
+%
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2010 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.
 %%
 
 clear;
--- a/examples/Automatic Music Transcription/SMALL_AMT_KSVD_Sparsity_test.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Automatic Music Transcription/SMALL_AMT_KSVD_Sparsity_test.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,14 +1,10 @@
-%%  DICTIONARY LEARNING FOR AUTOMATIC MUSIC TRANSCRIPTION EXAMPLE 1
+%%  Dictionary Learning for Automatic Music Transcription - KSVD sparsity
+%%  test
+%   
+%   *WARNING!* You should have SPAMS in your search path in order for this
+%   script to work.Due to licensing issues SPAMS can not be automatically 
+%   provided in SMALLbox (http://www.di.ens.fr/willow/SPAMS/downloads.html).
 %
-%   Centre for Digital Music, Queen Mary, University of London.
-%   This file copyright 2010 Ivan Damnjanovic.
-%
-%   This program is free software; you can redistribute it and/or
-%   modify it under the terms of the GNU General Public License as
-%   published by the Free Software Foundation; either version 2 of the
-%   License, or (at your option) any later version.  See the file
-%   COPYING included with this distribution for more information.
-%   
 %   This file contains an example of how SMALLbox can be used to test diferent
 %   dictionary learning techniques in Automatic Music Transcription problem.
 %   It calls generateAMT_Learning_Problem that will let you to choose midi,
@@ -19,6 +15,16 @@
 %   Problem.p notes spectrograms from training set Problem.b using
 %   dictionary learning technique defined in DL structure.
 %
+
+%
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2010 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.
 %%
 
 clear;
--- a/examples/Automatic Music Transcription/SMALL_AMT_SPAMS_test.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Automatic Music Transcription/SMALL_AMT_SPAMS_test.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,10 @@
-%%  DICTIONARY LEARNING FOR AUTOMATIC MUSIC TRANSCRIPTION EXAMPLE 1
+%%  Dictionary Learning for Automatic Music Transcription - SPAMS lambda
+%%  test
 %   
+%   *WARNING!* You should have SPAMS in your search path in order for this
+%   script to work.Due to licensing issues SPAMS can not be automatically 
+%   provided in SMALLbox (http://www.di.ens.fr/willow/SPAMS/downloads.html).
+%
 %   This file contains an example of how SMALLbox can be used to test diferent
 %   dictionary learning techniques in Automatic Music Transcription problem.
 %   It calls generateAMT_Learning_Problem that will let you to choose midi,
--- a/examples/Image Denoising/SMALL_ImgDenoise_DL_test_KSVDvsRLSDLA.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Image Denoising/SMALL_ImgDenoise_DL_test_KSVDvsRLSDLA.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,4 +1,5 @@
- %% DICTIONARY LEARNING FOR IMAGE DENOISING
+%%  Dictionary Learning for Image Denoising - KSVD vs Recursive Least Squares
+%
 %   This file contains an example of how SMALLbox can be used to test different
 %   dictionary learning techniques in Image Denoising problem.
 %   It calls generateImageDenoiseProblem that will let you to choose image,
--- a/examples/Image Denoising/SMALL_ImgDenoise_DL_test_KSVDvsSPAMS.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Image Denoising/SMALL_ImgDenoise_DL_test_KSVDvsSPAMS.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,6 +1,9 @@
-%% DICTIONARY LEARNING FOR IMAGE DENOISING
-
-%   
+%%  Dictionary Learning for Image Denoising - KSVD vs KSVDS vs SPAMS
+%
+%   *WARNING!* You should have SPAMS in your search path in order for this
+%   script to work.Due to licensing issues SPAMS can not be automatically 
+%   provided in SMALLbox (http://www.di.ens.fr/willow/SPAMS/downloads.html).
+%
 %   This file contains an example of how SMALLbox can be used to test different
 %   dictionary learning techniques in Image Denoising problem.
 %   It calls generateImageDenoiseProblem that will let you to choose image,
--- a/examples/Image Denoising/SMALL_ImgDenoise_DL_test_SPAMS_lambda.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Image Denoising/SMALL_ImgDenoise_DL_test_SPAMS_lambda.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,6 +1,9 @@
-%% DICTIONARY LEARNING FOR IMAGE DENOISING
+%%  Dictionary Learning for Image Denoising - SPAMS parameter test
 %
-%   
+%   *WARNING!* You should have SPAMS in your search path in order for this
+%   script to work.Due to licensing issues SPAMS can not be automatically 
+%   provided in SMALLbox (http://www.di.ens.fr/willow/SPAMS/downloads.html).
+%  
 %   This file contains an example of how SMALLbox can be used to test different
 %   dictionary learning techniques in Image Denoising problem.
 %   This example can be used to test SPAMS for different values of
--- a/examples/Image Denoising/SMALL_ImgDenoise_DL_test_Training_size.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Image Denoising/SMALL_ImgDenoise_DL_test_Training_size.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,6 +1,10 @@
-%% DICTIONARY LEARNING FOR IMAGE DENOISING
+%%  Dictionary Learning for Image Denoising - KSVD vs SPAMS Training size
+%%  test
 %
-   
+%   *WARNING!* You should have SPAMS in your search path in order for this
+%   script to work.Due to licensing issues SPAMS can not be automatically 
+%   provided in SMALLbox (http://www.di.ens.fr/willow/SPAMS/downloads.html).
+%
 %   This file contains an example of how SMALLbox can be used to test different
 %   dictionary learning techniques in Image Denoising problem.
 %   It calls generateImageDenoiseProblem that will let you to choose image,
--- a/examples/Pierre Villars/Pierre_Villars_Example.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/Pierre Villars/Pierre_Villars_Example.m	Mon Jun 13 14:58:00 2011 +0100
@@ -23,7 +23,7 @@
 
 %   Defining the Problem structure
 
-SMALL.Problem = Pierre_Problem();
+SMALL.Problem = generatePierre_Problem();
 
 %   Show original image and image that is used as a dictionary
 figure('Name', 'Original and Dictionary Image');
--- a/examples/SMALL_solver_test.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/SMALL_solver_test.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,5 @@
 function SMALL_solver_test
-%%   Example test of solvers from different toolboxes on Sparco problem 6
+%%  Example test of solvers from different toolboxes on Sparco problem 6
 %
 %   The main purpose of this example is to show how to use SMALL structure
 %   to solve SPARCO compressed sensing problems (1-11) and compare results
--- a/examples/SMALL_solver_test_Audio.m	Wed May 25 15:34:37 2011 +0100
+++ b/examples/SMALL_solver_test_Audio.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,15 +1,5 @@
 function SMALL_solver_test_Audio
-%%% Example test of solvers on Sparco audio source separation problems
-%
-%   Centre for Digital Music, Queen Mary, University of London.
-%   This file copyright 2009 Ivan Damnjanovic.
-%
-%   This program is free software; you can redistribute it and/or
-%   modify it under the terms of the GNU General Public License as
-%   published by the Free Software Foundation; either version 2 of the
-%   License, or (at your option) any later version.  See the file
-%   COPYING included with this distribution for more information.
-%   
+%%  Example test of solvers on Sparco audio source separation problems
 % 
 %   The main purpose of this example is to show how to use SMALL structure
 %   to solve SPARCO audio source3 separation problems (401-402) and  to 
@@ -44,8 +34,16 @@
 %   sources against original audio sources.
 %   SMALL_playAudio function plays audio sources of original and
 %   reconstructed signal as well as mixed signal.
-%   
-%   
+   
+%
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2009 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.   
 %%
 
 fprintf('\n\nExample test of solvers on Sparco Audio problems (401,402).\n\n');
--- a/solvers/SMALL_MP.m	Wed May 25 15:34:37 2011 +0100
+++ b/solvers/SMALL_MP.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,17 @@
 function [A]=SMALL_MP(Dict,X, m, maxNumCoef, errorGoal, varargin) 
-%%
+%%  Implementation of Matching Pursuit algorithm
+%   Sparse coding of a group of signals based on a given 
+%   dictionary and specified number of atoms to use. 
+%   input arguments: Dict - the dictionary
+%                   X - the signals to represent
+%                   m - number of atoms in Dictionary
+%                   errorGoal - the maximal allowed representation error for
+%                   each signal.
+%
+%   optional:         if Dict is function handle then Transpose Dictionary
+%                   handle needs to be specified.
+%
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -10,18 +22,6 @@
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
 %   
-% Sparse coding of a group of signals based on a given 
-% dictionary and specified number of atoms to use. 
-%  input arguments: Dict - the dictionary
-%                   X - the signals to represent
-%                   m - number of atoms in Dictionary
-%                   errorGoal - the maximal allowed representation error for
-%                   each signal.
-%
-% optional:         if Dict is function handle then Transpose Dictionary
-%                   handle needs to be specified.
-%
-
 %%
 % This Dictionary check is based on Thomas Blumensath work in sparsify 0_4 greedy solvers 
 explicitD=0;
--- a/solvers/SMALL_cgp.m	Wed May 25 15:34:37 2011 +0100
+++ b/solvers/SMALL_cgp.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,19 @@
 function [A, resF]=SMALL_cgp(Dict,X, m,  maxNumCoef, errorGoal, varargin) 
-%%% Conjugate Gradient Pursuit Multiple vectors sparse representation
+%%  Partial Conjugate Gradient Pursuit Multiple vectors sparse representation
+%   
+%   Sparse coding of a group of signals based on a given 
+%   dictionary and specified number of atoms to use. 
+%   input arguments: Dict - the dictionary
+%                  X - the signals to represent
+%                  m - number of atoms in Dictionary
+%                  errorGoal - the maximal allowed representation error for
+%                  each signal.
+%
+%   optional:        if Dict is function handle then Transpose Dictionary
+%                  handle needs to be specified.
+%
+%   output arguments: A - sparse coefficient matrix.
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -9,19 +23,6 @@
 %   published by the Free Software Foundation; either version 2 of the
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
-%   
-% Sparse coding of a group of signals based on a given 
-% dictionary and specified number of atoms to use. 
-% input arguments: Dict - the dictionary
-%                  X - the signals to represent
-%                  m - number of atoms in Dictionary
-%                  errorGoal - the maximal allowed representation error for
-%                  each signal.
-%
-% optional:        if Dict is function handle then Transpose Dictionary
-%                  handle needs to be specified.
-%
-% output arguments: A - sparse coefficient matrix.
 %
 %%
 
--- a/solvers/SMALL_chol.m	Wed May 25 15:34:37 2011 +0100
+++ b/solvers/SMALL_chol.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,19 @@
 function [A]=SMALL_chol(Dict,X, m, maxNumCoef, errorGoal, varargin) 
-%%
+%% Implementation of OMP with Cholesky factorisation
+% Sparse coding of a group of signals based on a given 
+% dictionary and specified number of atoms to use. 
+% input arguments:  Dict - the dictionary
+%                   X - the signals to represent
+%                   m - number of atoms in Dictionary
+%                   errorGoal - the maximal allowed representation error for
+%                   each signal.
+%
+% optional:         if Dict is function handle then Transpose Dictionary
+%                   handle needs to be specified.
+%
+% output arguments: A - sparse coefficient matrix.
+%
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -10,19 +24,6 @@
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
 %   
-% Sparse coding of a group of signals based on a given 
-% dictionary and specified number of atoms to use. 
-% input arguments: Dict - the dictionary
-%                   X - the signals to represent
-%                   m - number of atoms in Dictionary
-%                   errorGoal - the maximal allowed representation error for
-%                   each signal.
-%
-% optional:         if Dict is function handle then Transpose Dictionary
-%                   handle needs to be specified.
-%
-% output arguments: A - sparse coefficient matrix.
-%
 %%
 % This Dictionary check is based on Thomas Blumensath work in sparsify 0_4 greedy solvers 
 explicitD=0;
--- a/solvers/my_dummy_solver.m	Wed May 25 15:34:37 2011 +0100
+++ b/solvers/my_dummy_solver.m	Mon Jun 13 14:58:00 2011 +0100
@@ -0,0 +1,27 @@
+function [A]=my_dummy_solver(Dict,X, m,  maxNumCoef, errorGoal, varargin) 
+%%  Template function that can be used for solver implementation 
+%   
+%   Sparse coding of a group of signals based on a given 
+%   dictionary and specified number of atoms to use. 
+%
+%   input arguments:    Dict - the dictionary
+%                       X - the signals to represent
+%                       m - number of atoms in Dictionary
+%                       errorGoal - the maximal allowed representation
+%                                   error for each signal.
+%
+%   optional:      if Dict is function handle then Transpose Dictionary
+%                  handle needs to be specified.
+%
+%   output arguments: A - sparse coefficient matrix.
+
+%%  Change copyright notice as appropriate:
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2009 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.
+%%
--- a/util/SL_A.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SL_A.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,8 @@
 function y = SL_A(A, mode, m, n, x, I, dim)
-%%% Sparco to SparseLab implicit matix conversion
+%% Sparco to SparseLab implicit matix conversion
+%   This is auxilary function to allow implicit matrices from SPARCO
+%   to be used with SparsLab solvers
+
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
 %
@@ -8,9 +11,7 @@
 %   published by the Free Software Foundation; either version 2 of the
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
-%
-%   This is auxilary function to allow implicit matrices from SPARCO
-%   to be used with SparsLab solvers
+%%
 
 if (mode == 1)
     
--- a/util/SMALL_AMT_plot.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_AMT_plot.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,7 @@
 function figAMT=SMALL_AMT_plot(SMALL, AMT_res)
+%%   Function gets as input SMALL structure and plots AMT results
 %
+
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
 %
@@ -9,7 +11,7 @@
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
 %   
-%   Function gets as input SMALL structure and plots  AMT
+
 
 
 figAMT=figure('Name','Automatic Music Transcription');
--- a/util/SMALL_AudioDeNoiseResult.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_AudioDeNoiseResult.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,6 @@
 function SMALL_AudioDeNoiseResult(SMALL)
-%
+%% Plots the results of Audio denoising experiment - underconstruction
+
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
 %
--- a/util/SMALL_denoise.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_denoise.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,14 @@
 function solver=SMALL_denoise(Problem, solver)
-%%% SMALL denoising
+%% SMALL denoising - old replaced denoising function, would go out in v2.0
+%
+%   Function gets as input SMALL structure that contains SPARCO problem to
+%   be solved, name of the toolbox and solver, and parameters file for
+%   particular solver.
+%   It is based on omp/omps denoising by Ron Rubenstein (KSVD toolbox)
+%
+%   Outputs are denoised image, psnr, number of non-zero coeficients and 
+%   time spent
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -10,13 +19,7 @@
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
 %   
-%   Function gets as input SMALL structure that contains SPARCO problem to
-%   be solved, name of the toolbox and solver, and parameters file for
-%   particular solver.
-%   It is based on omp/omps denoising by Ron Rubenstein (KSVD toolbox)
-%
-%   Outputs are denoised image, psnr, number of non-zero coeficients and 
-%   time spent
+
 %%
   
   %%%%%  denoise the signal  %%%%%
--- a/util/SMALL_init_DL.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_init_DL.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,4 +1,10 @@
 function DL = SMALL_init_DL(varargin)
+%%   Function initialise SMALL structure for Dictionary Learning.
+%   Optional input variables:
+%       toolbox - name of Dictionary Learning toolbox you want to use
+%       name    - name of the algorithm from above toolbox
+%       param   - parameters you want to set
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2010 Ivan Damnjanovic.
@@ -9,11 +15,6 @@
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
 %
-%   Function initialise SMALL structure for Dictionary Learning.
-%   Optional input variables:
-%       toolbox - name of Dictionary Learning toolbox you want to use
-%       name    - name of the algorithm from above toolbox
-%       param   - parameters you want to set
 %%
 
 DL.toolbox=[];
--- a/util/SMALL_init_solver.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_init_solver.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,4 +1,10 @@
 function solver = SMALL_init_solver(varargin)
+%%   Function initialise SMALL structure for sparse representation.
+%   Optional input variables:
+%       toolbox - name of Dictionary Learning toolbox you want to use
+%       name    - name of the algorithm from above toolbox
+%       param   - parameters you want to set
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2010 Ivan Damnjanovic.
@@ -9,11 +15,6 @@
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
 %
-%   Function initialise SMALL structure for Dictionary Learning.
-%   Optional input variables:
-%       toolbox - name of Dictionary Learning toolbox you want to use
-%       name    - name of the algorithm from above toolbox
-%       param   - parameters you want to set
 %%
 
 solver.toolbox=[];
--- a/util/SMALL_midiGenerate.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_midiGenerate.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,13 @@
 function reconstructed=SMALL_midiGenerate(V, Problem)
-%%% Reconstraction of midi file from representation in the given dictionary
+%% Reconstraction of midi file from representation in the given dictionary
+%   
+%   SMALL_midiGenerate is a part of SMALLbox and can be use to reconstruct
+%   a midi file given representation of the training set (V) in the
+%   dictionary Problem.A.
+%   Output is reconstructed structure with two fields:
+%   -   reconstructed.notes - matrix with transcribed notes
+%   -   reconstructed.midi - midi representation of transcription
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -9,14 +17,6 @@
 %   published by the Free Software Foundation; either version 2 of the
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
-%   
-%   SMALL_midiGenerate is a part of SMALLbox and can be use to reconstruct
-%   a midi file given representation of the training set (V) in the
-%   dictionary Problem.A.
-%   Output is reconstructed structure with two fields:
-%   -   reconstructed.notes - matrix with transcribed notes
-%   -   reconstructed.midi - midi representation of transcription
-
 %%
 U=Problem.A;    %   Dictionary used for representation
 fs=Problem.fs;  %   Sampling rate
--- a/util/SMALL_playAudio.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_playAudio.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,8 @@
 function SMALL_playAudio(SMALL)
-%%% 
+%%  Plays reconstructed and original signals
+%   Function gets as input SMALL structure and play the original and
+%   the reconstructed signal
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -9,9 +12,8 @@
 %   published by the Free Software Foundation; either version 2 of the
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
-%   
-%   Function gets as input SMALL structure and play the original and
-%   the reconstructed signal
+%%   
+
 
 
 SMALL.solver.reconstructed  = SMALL.Problem.reconstruct(SMALL.solver.solution);
--- a/util/SMALL_plot.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_plot.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,8 @@
 function SMALL_plot(SMALL)
-%%% Plots coefficients and reconstructed signals 
+%% Plots coefficients and reconstructed signals 
+%   Function gets as input SMALL structure and plots  the solution and
+%   reconstructed signal
+
 %
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2009 Ivan Damnjanovic.
@@ -9,9 +12,7 @@
 %   published by the Free Software Foundation; either version 2 of the
 %   License, or (at your option) any later version.  See the file
 %   COPYING included with this distribution for more information.
-%   
-%   Function gets as input SMALL structure and plots  the solution and
-%   reconstructed signal
+%%   
 
   figure;
  
--- a/util/SMALL_showdict.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_showdict.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,7 +1,7 @@
 function x = SMALL_showdict(D,sz,n,m,varargin)
 %% SMALL_SHOWDICT Display a dictionary of image patches.
-%%  Reimplementation of showdict function from KSVD toolbox with Image
-%%  Processing toolbox dependecies removed
+%  Reimplementation of showdict function from KSVD toolbox with Image
+%  Processing toolbox dependecies removed
 %
 %  SMALL_SHOWDICT(D,SZ,N,M) displays the contents of the dictionary D, whos
 %  columns are 2-D image patches (in column-major order). SZ = [SX SY] is
@@ -25,7 +25,6 @@
 %  X = SMALL_SHOWDICT(...) returns a bitmat of the dictionary image without
 %  displaying the figure.
 
-
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2011 Ivan Damnjanovic.
 %
--- a/util/SMALL_solve.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_solve.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,5 @@
 function solver = SMALL_solve(Problem, solver)
-%%% SMALL sparse solver
+%% SMALL sparse solver
 %
 %   Function gets as input SMALL structure that contains SPARCO problem to
 %   be solved, name of the toolbox and solver, and parameters file for
--- a/util/SMALL_ssim_index.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_ssim_index.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,5 +1,5 @@
 function [mssim, ssim_map] = SMALL_ssim_index(img1, img2, K, window, L)
-
+%% SSIM Index - Removed dependence on Image Processing Toolbox 
 %========================================================================
 %SSIM Index, Version 1.0
 %Copyright(c) 2003 Zhou Wang
@@ -70,7 +70,7 @@
 %   imshow(max(0, ssim_map).^4)  %Shows the SSIM index map
 %
 
-% Removed dependency on Image Processing Toolbox - 
+% Removed dependence on Image Processing Toolbox - 
 %   Centre for Digital Music, Queen Mary, University of London.
 %   This file copyright 2011 Ivan Damnjanovic.
 %========================================================================
--- a/util/SMALL_swipe.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_swipe.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,20 +1,12 @@
 function [p,s] = SMALL_swipe(X,fs, f, plim,dlog2p,dERBs,woverlap,sTHR)
-% 
-%
-%   Centre for Digital Music, Queen Mary, University of London.
-%   This file copyright 2009 Ivan Damnjanovic.
-%
-%   This program is free software; you can redistribute it and/or
-%   modify it under the terms of the GNU General Public License as
-%   published by the Free Software Foundation; either version 2 of the
-%   License, or (at your option) any later version.  See the file
-%   COPYING included with this distribution for more information.
-%   
-%
+%%  Modified SWIPEP Pitch estimation using SWIPE'.
 %   This is modified swipep MATLAB code that is working directly in spectral
 %   domain and uses only one window size. The results are suboptimal
 %   comparing to original code. It is also converted to SWIPE which uses all
-%   the harmonics of the signal. 
+%   the harmonics of the signal.
+
+%   
+% 
 %
 %   SWIPEP Pitch estimation using SWIPE'.
 %    P = SWIPEP(X,Fs,[PMIN PMAX],DT,DLOG2P,DERBS,STHR) estimates the pitch 
@@ -70,6 +62,17 @@
 %    inspired pitch estimator for speech and music," J. Acoust. Soc. Am.
 %    124, 1638-1652.
 
+%
+%   Centre for Digital Music, Queen Mary, University of London.
+%   This file copyright 2009 Ivan Damnjanovic.
+%
+%   This program is free software; you can redistribute it and/or
+%   modify it under the terms of the GNU General Public License as
+%   published by the Free Software Foundation; either version 2 of the
+%   License, or (at your option) any later version.  See the file
+%   COPYING included with this distribution for more information.
+%%
+
 if ~ exist( 'plim', 'var' ) || isempty(plim), plim = [30 5000]; end
 %if ~ exist( 'dt', 'var' ) || isempty(dt), dt = 0.001; end
 if ~ exist( 'dlog2p', 'var' ) || isempty(dlog2p), dlog2p = 1/48; end
--- a/util/SMALL_vmrse_type2.m	Wed May 25 15:34:37 2011 +0100
+++ b/util/SMALL_vmrse_type2.m	Mon Jun 13 14:58:00 2011 +0100
@@ -1,6 +1,5 @@
 function [RMSErn, RMSEcd, rn_im, cd_im] = SMALL_vmrse_type2(orig, corr, recon)
-%%   Implementation of VectorRMSE type2
-%
+%%  Implementation of VectorRMSE type2
 % 
 %   Input:
 %       - Original image