From 9b518922e75e1e79ca70967680c32ffce6f65df9 Mon Sep 17 00:00:00 2001
From: Julien Malik <julien.malik@c-s.fr>
Date: Sat, 28 Jan 2012 23:56:44 +0100
Subject: [PATCH] DOC: Dimension reduction examples improvements
---
Code/BasicFilters/otbMNFImageFilter.h | 8 ++---
Examples/DimensionReduction/ICAExample.cxx | 7 +++--
Examples/DimensionReduction/MNFExample.cxx | 23 +++++++++++----
.../MaximumAutocorrelationFactor.cxx | 2 +-
Examples/DimensionReduction/NAPCAExample.cxx | 29 +++++++++++++++++--
5 files changed, 53 insertions(+), 16 deletions(-)
diff --git a/Code/BasicFilters/otbMNFImageFilter.h b/Code/BasicFilters/otbMNFImageFilter.h
index e488ce1950..13dc7add7d 100644
--- a/Code/BasicFilters/otbMNFImageFilter.h
+++ b/Code/BasicFilters/otbMNFImageFilter.h
@@ -29,15 +29,15 @@
namespace otb {
/** \class MNFImageFilter
- * \brief Performs a Minimum Noise Fraction analysis of a vector image.
+ * \brief Performs a Maximum Noise Fraction analysis of a vector image.
*
* The internal structure of this filter is a filter-to-filter like structure.
- * The estimation of the covariance matrix has persistent capabilities...
+ * The estimation of the covariance matrix is streamed
*
* The high pass filter which has to be used for the noise estimation is templated
- * for a better scalability...
+ * for a better scalability.
*
- * TODO? Utiliser une 2e entree pour donner directement une image de bruit ??
+ * TODO? Utiliser une 2e entree pour donner directement une image de bruit ?
*
* \sa otbStreamingStatisticsVectorImageFilter
* \sa PCAImageFiler
diff --git a/Examples/DimensionReduction/ICAExample.cxx b/Examples/DimensionReduction/ICAExample.cxx
index bc3d10ae2f..7aab2fe9e6 100644
--- a/Examples/DimensionReduction/ICAExample.cxx
+++ b/Examples/DimensionReduction/ICAExample.cxx
@@ -33,10 +33,13 @@
// \doxygen{otb}{FastICAImageFilter}.
// This filter computes a Fast Independant Components Analysis transform.
//
-// Like Principal Components Analysis, FastICA computes a set of
+// Like Principal Components Analysis, Independent Component Analysis
+// \cite{jutten1991blind} computes a set of
// orthogonal linear combinations, but the criterion of Fast ICA is
// different: instead of maximizing variance, it tries to maximize
-// stastistical independance between components. In Fast ICA,
+// stastistical independance between components.
+//
+// In the Fast ICA algorithm \cite{hyvarinen1999fast},
// statistical independance is mesured by evaluating non-Gaussianity
// of the components, and the maximization is done in an iterative way.
diff --git a/Examples/DimensionReduction/MNFExample.cxx b/Examples/DimensionReduction/MNFExample.cxx
index 0d56269f08..466bc969f1 100644
--- a/Examples/DimensionReduction/MNFExample.cxx
+++ b/Examples/DimensionReduction/MNFExample.cxx
@@ -30,19 +30,19 @@
// Software Guide : BeginLatex
//
// This example illustrates the use of the
-// \doxygen{otb}{MNFImageFilter}. This filter computes a Minimum
-// Noise Fraction transform \cite{nielsen2011kernel} using an
+// \doxygen{otb}{MNFImageFilter}. This filter computes a Maximum
+// Noise Fraction transform \cite{green1988transformation} using an
// efficient method based on the inner product in order to compute the
// covariance matrix.
//
-// The Minimum Noise Fraction transform is a sequence of two Principal
-// Components Analysis transform. The first transform is based on an
+// The Maximum Noise Fraction transform is a sequence of two Principal
+// Component Analysis transforms. The first transform is based on an
// estimated covariance matrix of the noise, and intends to whiten the
// input image (noise with unit variance and no correlation between
// bands).
//
-// The second Principal Components Analysis is then applied to the
-// noise-whitened image, giving the Minimum Noise Fraction transform.
+// The second Principal Component Analysis is then applied to the
+// noise-whitened image, giving the Maximum Noise Fraction transform.
//
// In this implementation, noise is estimated from a local window.
//
@@ -106,6 +106,17 @@ int main(int argc, char* argv[])
// Software Guide : BeginLatex
//
+ // In contrast with standard Principal Component Analysis, MNF
+ // needs an estimation of the noise correlation matrix
+ // in the dataset prior to transformation.
+ //
+ // A classical approach is to use spatial gradient images
+ // and infer the noise correlation matrix from it.
+ // The method of noise estimation can be customized
+ // by templating the \doxygen{otb}{MNFImageFilter}
+ // with the desired noise estimation method.
+ //
+ // In this implementation, noise is estimated from a local window.
// We define the type of the noise filter.
//
// Software Guide : EndLatex
diff --git a/Examples/DimensionReduction/MaximumAutocorrelationFactor.cxx b/Examples/DimensionReduction/MaximumAutocorrelationFactor.cxx
index 2a96ec70c1..752826e4a7 100644
--- a/Examples/DimensionReduction/MaximumAutocorrelationFactor.cxx
+++ b/Examples/DimensionReduction/MaximumAutocorrelationFactor.cxx
@@ -31,7 +31,7 @@
// \doxygen{otb}{MaximumAutocorrelationFactorImageFilter}, which
// performs a Maximum Autocorrelation Factor transform \cite{nielsen2011kernel}. Like
// PCA, MAF tries to find a set of orthogonal linear transform, but
-// the criterion to maximize is the auto-correlation rather than the
+// the criterion to maximize is the spatial auto-correlation rather than the
// variance.
//
// Auto-correlation is the correlation between the component and a
diff --git a/Examples/DimensionReduction/NAPCAExample.cxx b/Examples/DimensionReduction/NAPCAExample.cxx
index 448b766e91..ea61405997 100644
--- a/Examples/DimensionReduction/NAPCAExample.cxx
+++ b/Examples/DimensionReduction/NAPCAExample.cxx
@@ -30,11 +30,22 @@
// Software Guide : BeginLatex
//
// This example illustrates the use of the
-// \doxygen{otb}{NAPCAImageFilter}.
-// This filter computes a Principal Component Analysis using an
+// \doxygen{otb}{NAPCAImageFilter}. This filter computes a Noise-Adjusted
+// Principal Component Analysis transform \cite{lee1990enhancement} using an
// efficient method based on the inner product in order to compute the
// covariance matrix.
//
+// The Noise-Adjusted Principal Component Analysis transform is a sequence
+// of two Principal Component Analysis transforms. The first transform is based
+// on an estimated covariance matrix of the noise, and intends to whiten the
+// input image (noise with unit variance and no correlation between
+// bands).
+//
+// The second Principal Component Analysis is then applied to the
+// noise-whitened image, giving the Maximum Noise Fraction transform.
+//
+// It is basically a reformulation of the Maximum Noise Fraction algorithm.
+//
// The first step required to use this filter is to include its header file.
//
// Software Guide : EndLatex
@@ -95,6 +106,17 @@ int main(int argc, char* argv[])
// Software Guide : BeginLatex
//
+ // In contrast with standard Principal Component Analysis, NA-PCA
+ // needs an estimation of the noise correlation matrix
+ // in the dataset prior to transformation.
+ //
+ // A classical approach is to use spatial gradient images
+ // and infer the noise correlation matrix from it.
+ // The method of noise estimation can be customized
+ // by templating the \doxygen{otb}{NAPCAImageFilter}
+ // with the desired noise estimation method.
+ //
+ // In this implementation, noise is estimated from a local window.
// We define the type of the noise filter.
//
// Software Guide : EndLatex
@@ -107,7 +129,8 @@ int main(int argc, char* argv[])
// Software Guide : BeginLatex
//
// We define the type for the filter. It is templated over the input
- // and the output image types and also the transformation direction. The
+ // and the output image types, the noise estimation filter type,
+ // and also the transformation direction. The
// internal structure of this filter is a filter-to-filter like structure.
// We can now the instantiate the filter.
//
--
GitLab