From 94fe997cba2bcf326d8204baef87a39ee9e0d6a2 Mon Sep 17 00:00:00 2001
From: Jordi Inglada <jordi.inglada@orfeo-toolbox.org>
Date: Fri, 30 Jun 2006 12:18:53 +0000
Subject: [PATCH] Mise a jour exemples
---
Latex/Insight.bib | 599 ++++++++++++++++++
.../Art/BayesianPluginClassifier.fig | 260 ++++++++
SoftwareGuide/Art/DudaClassifier.fig | 575 +++++++++++++++++
.../StatisticalClassificationFramework.fig | 116 ++++
.../Art/TwoNormalDensityFunctionPlot.jpg | Bin 0 -> 20283 bytes
SoftwareGuide/Examples/CMakeLists.txt | 11 +-
SoftwareGuide/Latex/ChangeDetection.tex | 41 +-
SoftwareGuide/Latex/Classification.tex | 210 +++++-
SoftwareGuide/Latex/SoftwareGuide.tex | 1 +
SoftwareGuide/Latex/Visualization.tex | 27 +-
10 files changed, 1804 insertions(+), 36 deletions(-)
create mode 100755 SoftwareGuide/Art/BayesianPluginClassifier.fig
create mode 100755 SoftwareGuide/Art/DudaClassifier.fig
create mode 100755 SoftwareGuide/Art/StatisticalClassificationFramework.fig
create mode 100755 SoftwareGuide/Art/TwoNormalDensityFunctionPlot.jpg
diff --git a/Latex/Insight.bib b/Latex/Insight.bib
index f904d83ba6..fd425245a8 100644
--- a/Latex/Insight.bib
+++ b/Latex/Insight.bib
@@ -10124,3 +10124,602 @@ number = 10,
pages = "2235--2243",
year = 2001,
}
+
+
+@article{faces,
+author = {Ashok Samal and Prasana A. Iyengar},
+title = {Automatic recognition and analysis of human faces and facial
+expressions: a survey},
+journal = {Pattern Recogn.},
+volume = {25},
+number = {1},
+year = {1992},
+issn = {0031-3203},
+pages = {65--77},
+publisher = {Elsevier Science Inc.},
+ }
+
+
+@article{handwritten,
+author = {R. H. Davis and J. Lyall},
+title = {Recognition of handwritten characters: a review},
+journal = {Image Vision Comput.},
+volume = {4},
+number = {4},
+year = {1986},
+issn = {0262-8856},
+pages = {208--218},
+publisher = {Butterworth-Heinemann},
+ }
+
+%%%%%%%%%%%%%%%% Comprehension %%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+
+@ARTICLE{prince_intel,
+AUTHOR = "V. Prince",
+TITLE="Expertise naturelle, expertise artificielle, vers quels paradigmes cognitifs?",
+JOURNAL="Intellectica - Expertise et Sciences Cognitives",
+YEAR=1991,
+NUMBER=12,
+PAGES = "7--31",
+}
+
+@ARTICLE{kriv_intel,
+AUTHOR = "J.P. Krivine and J.M. David",
+TITLE="L'acquisition des connaissances vue comme un processus de modélisation : méthodes et outils.",
+JOURNAL="Intellectica - Expertise et Sciences Cognitives",
+YEAR=1991,
+NUMBER=12,
+PAGES = "101--137",
+}
+
+@inproceedings{ pfefferspook,
+ author = "Avi Pfeffer and Daphne Koller and Brian Milch and Ken T. Takusagawa",
+ title = "{SPOOK}: {A} system for probabilistic object-oriented knowledge representation",
+ pages = "541--550",
+ url = "citeseer.nj.nec.com/pfeffer99spook.html" }
+
+@inproceedings{ koller97objectoriented,
+ author = "Daphne Koller and Avi Pfeffer",
+ title = "Object-Oriented Bayesian Networks",
+ booktitle = "Proceedings of the Thirteenth Conference on Uncertainty in Artificial Intelligence ({UAI}-97)",
+ pages = "302--313",
+ year = "1997",
+ url = "citeseer.nj.nec.com/koller97objectoriented.html" }
+
+
+@PHDTHESIS{these_moissinac,
+AUTHOR = "Henri Moissinac Massénat",
+TITLE = "Utilisation conjointe d'informations symboliques et de mesures numériques dans la prise de décision en traitement d'images",
+SCHOOL = "Ecole Nationale Supérieure des Télécommunications, Paris",
+YEAR = 1996,
+}
+
+%%%%%%%%%%%%%%%%%%%%%% Indexation %%%%%%%%%%%%%%%%%%%%%%%
+
+%%%%%%%%%%%%%%%%%%%%%% Gestalt %%%%%%%%%%%%%%%%%%%%%%%%%%
+
+@article{desolneux2000,
+author = {Agnès Desolneux and Lionel Moisan and Jean-Michel Morel},
+title = "{Edge Detection by Helmholtz Principle}",
+journal = {J. Math. Imaging Vis.},
+volume = {14},
+number = {3},
+year = {2001},
+issn = {0924-9907},
+pages = {271--284},
+doi = {http://dx.doi.org/10.1023/A:1011290230196},
+publisher = {Kluwer Academic Publishers},
+}
+
+
+%%%%%%%%%%%%%%%%%%%%% Objets %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+@misc{ jaimes00integrating,
+ author = "A. Jaimes and S. Chang",
+ title = "Integrating Multiple Classifiers in Visual Object Detectors Learned from
+ User Input",
+ text = "A. Jaimes and S.-F. Chang, Integrating Multiple Classifiers in Visual Object
+ Detectors Learned from User Input, Invited paper, session on Image and Video
+ Databases, 4th Asian Conference on Computer Vision (ACCV 2000), Taipei,
+ Taiwan, January 8-11, 2000.",
+ year = "2000",
+ url = "citeseer.nj.nec.com/jaimes00integrating.html" }
+
+
+@techreport{ liu98deformable,
+ author = "Lifeng Liu and Stan Sclaroff",
+ title = "Deformable Shape Detection and Description via Model-Based Region Grouping",
+ number = "98-017",
+ month = "4,",
+ year = "1998",
+ url = "citeseer.nj.nec.com/article/liu98deformable.html" }
+
+@techreport{ vetter97bootstrapping,
+ author = "Thomas Vetter and Michael J. Jones and Tomaso Poggio",
+ title = "A Bootstrapping Algorithm for Learning Linear Models of Object Classes",
+ number = "AIM-1600",
+ pages = "12",
+ year = "1997",
+ url = "citeseer.nj.nec.com/vetter97bootstrapping.html" }
+
+@techreport{ jones96modelbased,
+ author = "Michael J. Jones and Tomaso Poggio",
+ title = "Model-Based Matching by Linear Combinations of Prototypes",
+ number = "AIM-1583",
+ pages = "33",
+ year = "1996",
+ url = "citeseer.nj.nec.com/jones96modelbased.html" }
+
+@inproceedings{gfod,
+AUTHOR="Constanine P. Papageorgiou and Michael Oren and Tomaso Poggio",
+TITLE = "{A General Framework for Object Detection}",
+BOOKTITLE = "Intenational Conference on Computer Vision, Bombay, India",
+MONTH=jan,
+YEAR = 1998,
+}
+@article{ pontil98support,
+ author = "Massimiliano Pontil and Alessandro Verri",
+ title = "Support Vector Machines for 3D Object Recognition",
+ journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
+ volume = "20",
+ number = "6",
+ pages = "637-646",
+ year = "1998",
+ url = "citeseer.nj.nec.com/pontil98support.html" }
+
+@inproceedings{cbop,
+AUTHOR = "B. Heisele and T. Serre and M. Pontil and T. Vetter and T. Poggio",
+TITLE= "{Categorization by Learning and Combining Object Parts}",
+BOOKTITLE="Advances in Neural Information Processing Systems (NIPS'01), Vancouver, Canada",
+YEAR = 2001,
+}
+
+%%%%%%%%%%%%%%%%%%%%% Raisonnement spatial %%%%%%%%%%%%%%%
+
+@misc{ sentences_images,
+ author = "Alicia Abella and John R. Kender",
+ title = "{From Images to Sentences via Spatial Relations}",
+ text = "AT\&T Shannon Laboratories and Columbia University",
+ url = "www.cs.columbia.edu/~jrk/research/speimg.ps" }
+
+
+
+%%%%%%%%%%%%%%%%%%%%% INVARIANTS %%%%%%%%%%%%%%%%%%%%%%%%%%
+
+@misc{cllis,
+author="Aleksandra Mojsilovic and Bernice Rogowitz",
+title="Capturing image semantics with low-level descriptors",
+text="IBM TJ Watson Research Center, Hawthorne, NY, 10532",
+url = "www.research.ibm.com/visualanalysis/papers/mojsilovic.pdf",
+}
+
+
+
+
+@techreport{nonparam_inria,
+AUTHOR = "Zhong-Dan Lan and Roger Mohr",
+TITLE = "{Non-parametric Invariants and Application to Matching}",
+INSTITUTION = "INRIA, no. 3246",
+MONTH = sep,
+YEAR = 1997,
+}
+
+@techreport{ rothe94general,
+ author = "Irene Rothe and Klaus Voss and Herbert S{\"u}sse and J{\"o}rg Rothe",
+ title = "A General Method to Determine Invariants",
+ number = "TR503",
+ year = "1994",
+ url = "citeseer.nj.nec.com/rothe94general.html" }
+
+
+
+@techreport{questions_inria,
+AUTHOR = "Donald Geman and Bruno Jedynak",
+TITLE = "{Shape Recognition and Twenty Questions}",
+INSTITUTION = "INRIA, no. 2155",
+MONTH = nov,
+YEAR = 1993,
+}
+
+
+@ARTICLE{mostafa,
+AUTHOR="Y.S. Abu-Mostafa and D. Psaltis",
+TITLE="{Recoginitive aspects of moments invariants}",
+JOURNAL="IEEE Transactions on Pattern Analysis and Machine Intelligence",
+VOLUME=6,
+YEAR=1984,
+PAGES="698--706",
+}
+
+
+%%%%%%%%%%%%%%%%%% Perception %%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+@article{mltmfe,
+author="Y. Machrouh and J.-S. Liénard and P. Tarroux",
+title="Multiscale feature extraction from the visual environment in an active vision system",
+journal="Lecture Notes in Computer Science",
+volume=2059,
+pages="388--397",
+year=2001
+}
+
+@misc{mdsc,
+author="Jean Petitot",
+url="http://www.crea.polytechnique.fr/JeanPetitot/JPmodeles.html",
+}
+
+@article{ jaimes01learning,
+ author = "Alejandro Jaimes and Shih-Fu Chang",
+ title = "Learning Structured Visual Detectors from User Input at Multiple Levels",
+ journal = "International Journal of Image and Graphics",
+ volume = "1",
+ number = "3",
+ pages = "415-444",
+ year = "2001",
+ url = "citeseer.nj.nec.com/484241.html" }
+
+%%%%%%%%%%%%%%%%%% Pattern Matching %%%%%%%%%%%%%%%%%%%%%%
+@misc{ hagedoorn97reliable,
+ author = "M. Hagedoorn and R. Veltkamp",
+ title = "Reliable and efficient pattern matching using an affine invariant metric",
+ text = "M. Hagedoorn and R. C. Veltkamp. Reliable and efficient pattern matching
+ using an affine invariant metric. Technical Report RUU-CS-97-33, Dept. of
+ Computing Science, Utrecht University, The Netherlands, 1997.",
+ year = "1997",
+ url = "citeseer.nj.nec.com/article/hagedoorn97reliable.html" }
+
+
+@misc{ veltkamp99stateart,
+ author = "R. Veltkamp and M. Hagedoorn",
+ title = "State-of-the-art in shape matching",
+ text = "Remco C. Veltkamp and Michiel Hagedoorn. State-of-the-art in shape matching.
+ Technical Report UU-CS-1999-27, Utrecht University, the Netherlands, 1999.",
+ year = "1999",
+ url = "citeseer.nj.nec.com/veltkamp99stateart.html" }
+
+@techreport{ alt96discrete,
+ author = "Helmut Alt and Leonidas J. Guibas",
+ title = "Discrete Geometric Shapes: Matching, Interpolation, and Approximation {A} Survey",
+ number = "B 96-11",
+ year = "1996",
+ url = "citeseer.nj.nec.com/alt96discrete.html" }
+
+@article{ huttenlocher93comparing,
+ author = "D.~Huttenlocher and D.~Klanderman and A.~Rucklige",
+ title = "Comparing images using the {H}ausdorff distance",
+ journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
+ volume = "15",
+ number = "9",
+ month = "September",
+ pages = "850--863",
+ year = "1993",
+ url = "citeseer.nj.nec.com/huttenlocher93comparing.html" }
+
+
+@techreport{veltkamp,
+ author = "R. Veltkamp and M. Hagedoorn",
+ title = "State-of-the-art in shape matching",
+ number = "UU-CS-1999-27",
+ institution = "Utrecht University, the Netherlands",
+ year = "1999",
+ url = "citeseer.ist.psu.edu/veltkamp99stateart.html" }
+
+
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+@inproceedings{pope1,
+author = "Arthur A. Pope and David G. Lowe",
+title = "{Learning Appearance Models for Object Recognition}",
+booktitle = "International Workshop on Object Representation for Computer Vision, Cambridge, UK.",
+month = apr,
+year = 1996,
+}
+
+@article{hamdan,
+author = "R. Hamdan and F. Heitz and L. Thoraval",
+title = "{Mod\`eles probabilistes d'apparence : une repr\'esentation approch\'ee de faible complexit\'e}",
+journal = "Traitement du Signal",
+volume = 18,
+number = 3,
+year = 2001,
+}
+
+
+@Article{TurkPent1991,
+author = "Turk, Matthew and Pentland, Alex Paul",
+year = "1991",
+title = "Eigenfaces for Recognition.",
+journal = "Journal of Cognitive Neuroscience",
+volume = "3",
+number = "1",
+pages = "71--86",
+}
+
+@article{turkran,
+author = "M. Turk",
+title = "{A random walk through eigenspace}",
+journal = "IEICE Trans. Inf. Syst.",
+volume = "E84-D",
+number = 12,
+pages = "1586--1695",
+year = 2001,
+}
+
+@ARTICLE{lbica,
+author = {H. Le Borgne and A. Gu\'erin-Dugu\'e and A. Antoniadis},
+title = {Representation of images for classification with independent features},
+journal = {Pattern Recognition Letters},
+year = {2004},
+volume = {25},
+number = {2},
+pages = {141-154},
+month = {jan},
+}
+
+@ARTICLE{mclaughlin,
+AUTHOR="R.A. Laughlin and M.D Alder",
+TITLE="{The Hough transform versus the upwrite}",
+JOURNAL="IEEE Transactions on Pattern Analysis and Machine Intelligence",
+VOLUME=20,
+number=4,
+YEAR=1998,
+MONTH=apr,
+PAGES="396--400",
+}
+
+@article{yin,
+ author = {Peng-Yeng Yin},
+ title = {A new circle/ellipse detector using genetic algorithms},
+ journal = {Pattern Recogn. Lett.},
+ volume = {20},
+ number = {7},
+ year = {1999},
+ issn = {0167-8655},
+ pages = {731--740},
+ doi = {http://dx.doi.org/10.1016/S0167-8655(99)00037-9},
+ publisher = {Elsevier Science Inc.},
+ address = {New York, NY, USA},
+ }
+
+
+@phdthesis{these_derrode,
+author="St\'ephane Derrode",
+title="{Repr\'esentation de formes planes \`a niveaux de gris par diff\' erentes approximations de Fourier-Mellin analytique en vue d'indexation de bases d'images}",
+year=1999,
+school="Universit\'e de Rennes I",
+}
+
+
+@article{derrode,
+title="{Robust and efficient Fourier-Mellin transform approximations for gray-level image reconstruction and complete invariant description}",
+author="S. Derrode and F. Ghorbel",
+journal ="Computer Vision and Image Understanding",
+volume=83,
+number=1,
+pages="57--78",
+month=jul,
+year =2001,
+}
+
+@ARTICLE{Canny,
+AUTHOR="J.Canny",
+TITLE="{A computational approach to edge detection}",
+JOURNAL="IEEE Transactions on Pattern Analysis and Machine Intelligence",
+VOLUME=8,
+YEAR=1986,
+MONTH=apr,
+PAGES="79--698",
+}
+
+
+@BOOK{vapnik,
+AUTHOR="V. Vapnik",
+TITLE="Statistical learning theory",
+PUBLISHER="John Wiley and Sons, NewYork",
+YEAR=1998,
+}
+
+@TechReport{joachims,
+AUTHOR ="Thorsten Joachims",
+TITLE = "{Text Categorization with Support Vector Machines: Learning with Many Relevant Features}",
+INSTITUTION = "Computer Science of The University of dortmund",
+YEAR = 1997,
+MONTH =nov,
+}
+
+@misc{ osuna,
+ author = "E. Osuna and R. Freund and F. Girosi",
+ title = "Training support vector machines:an application to face detection",
+ text = "E. Osuna, R. Freund, and F. Girosi. Training support vector machines: an
+ application to face detection. CVPR'97, 1997.",
+ year = "1997",
+ url = "citeseer.nj.nec.com/osuna97training.html" }
+
+
+
+@ARTICLE{datcu98a,
+AUTHOR="Mihai Datcu and Klaus Seidel and Marc Walessa",
+TITLE="{Spatial Information Retrieval from Remote-Sensing Images --- Part I: Information Theoretical Perspective}",
+JOURNAL="IEEE Transactions Geoscience Remote Sensing",
+VOLUME=36,
+number = 5,
+month = sep,
+YEAR=1998,
+PAGES="1431--1445",
+}
+
+@ARTICLE{datcu98b,
+AUTHOR="Michael Schroder and Hubert Rehrauer and Klaus Seidel and Mihai Datcu",
+TITLE="{Spatial Information Retrieval from Remote-Sensing Images --- Part II: Gibbs-Markov Random Fields}",
+JOURNAL="IEEE Transactions Geoscience Remote Sensing",
+VOLUME=36,
+number = 5,
+month = sep,
+YEAR=1998,
+PAGES="1446--1455",
+}
+
+@ARTICLE{schroder00,
+AUTHOR="Michael Schroder and Hubert Rehrauer and Klaus Seidel and Mihai Datcu",
+TITLE="{Interactive Learning and Probabilistic Retrieval in Remote Sensing Image Archives}",
+JOURNAL="IEEE Transactions Geoscience Remote Sensing",
+VOLUME=38,
+number = 5,
+month = sep,
+YEAR=2000,
+PAGES="2288--2298",
+}
+
+@ARTICLE{dellacqua01,
+AUTHOR="Fabio Dell'Acqua and Paolo Gamba",
+TITLE="{Query-by-shape in Meteorological Image Archives Using the Point Diffusion Technique}",
+JOURNAL="IEEE Transactions Geoscience Remote Sensing",
+VOLUME=39,
+number = 9,
+month = sep,
+YEAR=2001,
+PAGES="1834--1843",
+}
+
+@ARTICLE{datcu03,
+AUTHOR="Mihai Datcu and Herbert Daschiel and Andrea Pelizzari and Marco Quartulli and Annalisa Galoppo and Andrea Colapicchioni and Marco Pastori and Klaus Seidel and Pier~Giorgio Marchetti",
+TITLE="{Information Mining in Remote Sensing Image Archives: System Concepts}",
+JOURNAL="IEEE Transactions Geoscience Remote Sensing",
+VOLUME=41,
+number = 12,
+month = dec,
+YEAR=2003,
+PAGES="2923--2936",
+}
+
+@ARTICLE{datcu05,
+AUTHOR="Herbert Daschiel and Mihai Datcu",
+TITLE="{Information Mining in Remote Sensing Image Archives: System Evaluation}",
+JOURNAL="IEEE Transactions Geoscience Remote Sensing",
+VOLUME=43,
+number = 1,
+month = jan,
+YEAR=2005,
+PAGES="188--199",
+}
+
+@inproceedings{bruzzoneSVM,
+author = "L. Bruzzone and F. Melgani",
+title = "Support vector machines for classification of hyperspectral remote-sensing images",
+booktitle = "IEEE International Geoscience and Remote Sensing Symposium, IGARSS",
+volume=1,
+pages = "506--508",
+month = jun,
+year = 2002,
+}
+
+@article{fastica,
+author="A. Hyv{\"a}rinen and E. Oja",
+title="A fast fixed-point algorithm for independent component analysis",
+journal="Neural Computation",
+volume=9,
+number=7,
+pages="1483--1492",
+year=1997,
+}
+
+@ARTICLE{guy,
+AUTHOR="S. Thorpe and A. Delorme and R. Van~Rullen",
+TITLE="{Spike-based strategies for rapid processing}",
+JOURNAL="Neural Networks",
+VOLUME=14,
+YEAR=2001,
+PAGES="715--725",
+}
+
+@article{burges,
+author = "C.J.C. Burges",
+title="{A Tutorial on Support Vector Machines for Pattern Recognition}",
+journal="Data Mining and Knowledge Discovery",
+volume= 2,
+number= 2,
+pages="121--167",
+year = 1998,
+}
+
+@INPROCEEDINGS{Rochery03a,
+ author = {M. Rochery and I. H. Jermyn and J. Zerubia},
+ month = oct,
+ year = 2003,
+ title = {Higher Order Active Contours and their Application to the Detection
+ of Line Networks in Satellite Imagery},
+ booktitle = "Proc. IEEE Workshop Variational, Geometric and Level Set Methods in Computer Vision",
+ address = {at ICCV, Nice, France},
+ abstract = {We present a novel method for the incorporation of shape
+ information into active contour models, and apply it to the
+ extraction of line networks (e.g. road, water) from satellite
+ imagery. The method is based on a new class of contour energies.
+ These energies are quadratic on the space of one-chains in the
+ image, as opposed to classical energies, which are linear. They
+ can be expressed as double integrals on the contour, and thus
+ incorporate non-trivial interactions between different contour
+ points. The new energies describe families of contours that share
+ complex geometric properties, without making reference to any
+ particular shape. Networks fall into such a family, and to model
+ them we make a particular choice of quadratic energy whose minima
+ are reticulated. To optimize the energies, we use a level set
+ approach. The forces derived from the new energies are non-local
+ however, thus necessitating an extension of standard level set
+ methods. Promising experimental results are obtained using real
+ images.}
+}
+@article{stoica,
+author = "R. Stoica and X. Descombes and J. Zerubia",
+title = "{A Gibbs point process for road extraction from remotely sensed images}",
+journal = "International Journal of Computer Vision",
+volume = 57,
+number = 2,
+pages = "121--136",
+year = 2004,
+}
+
+@inproceedings{joachims-errors,
+author = "T. Joachims",
+title = "{Estimating the Generalization Performance of a SVM Efficiently}",
+booktitle = "Proceedings of the International Conference on Machine Learning",
+publisher = "Morgan Kaufman",
+year = 2000,
+}
+
+@incollection{svmlight,
+author = "T. Joachims",
+title = "Making large-Scale SVM Learning Practical",
+booktitle = "Advances in Kernel Methods - Support Vector Learning",
+editor = "B. Schölkopf and C. Burges and A. Smola",
+publisher = "MIT-Press",
+year = 1999,
+}
+
+@misc{ hsu01comparison,
+ author = "C. Hsu and C. Lin",
+ title = "A comparison of methods for multi-class support vector machines",
+ text = "C.-W. Hsu and C.-J. Lin. A comparison of methods for multi-class support
+ vector machines. Technical report, Department of Computer Science and Information
+ Engineering, National Taiwan University, Taipei, Taiwan, 2001. 19",
+ year = "2001",
+ url = "citeseer.ist.psu.edu/hsu01comparison.html" }
+
+@misc{ weston98multiclass,
+ author = "J. Weston and C. Watkins",
+ title = "Multi-class support vector machines",
+ text = "J. Weston and C. Watkins. Multi-class support vector machines. Technical
+ Report CSD-TR-98-04, Department of Computer Science, Royal Holloway, University
+ of London, Egham, TW20 0EX, UK, 1998.",
+ year = "1998",
+ url = "citeseer.ist.psu.edu/weston98multiclass.html" }
+
+@inproceedings{ allwein00reducing,
+ author = "Erin L. Allwein and Robert E. Schapire and Yoram Singer",
+ title = "Reducing Multiclass to Binary: {A} Unifying Approach for Margin Classifiers",
+ booktitle = "Proc. 17th International Conf. on Machine Learning",
+ publisher = "Morgan Kaufmann, San Francisco, CA",
+ pages = "9--16",
+ year = "2000",
+ url = "citeseer.ist.psu.edu/allwein00reducing.html" }
+
+
diff --git a/SoftwareGuide/Art/BayesianPluginClassifier.fig b/SoftwareGuide/Art/BayesianPluginClassifier.fig
new file mode 100755
index 0000000000..196a5e3a66
--- /dev/null
+++ b/SoftwareGuide/Art/BayesianPluginClassifier.fig
@@ -0,0 +1,260 @@
+#FIG 3.2
+Landscape
+Center
+Inches
+Letter
+100.00
+Single
+-2
+1200 2
+0 32 #848284
+0 33 #c6c3c6
+0 34 #e7e3e7
+0 35 #848284
+0 36 #bdbebd
+0 37 #dedfde
+0 38 #8c8a8c
+0 39 #8c8e8c
+0 40 #8c8e8c
+0 41 #8c8a8c
+0 42 #4a454a
+0 43 #8c8a8c
+0 44 #848684
+0 45 #c6c7c6
+0 46 #e7e7e7
+0 47 #f7f7f7
+0 48 #9c9a9c
+0 49 #737573
+0 50 #424542
+0 51 #e7e3e7
+0 52 #c6c3c6
+0 53 #8c8a8c
+0 54 #8c8a8c
+0 55 #424142
+0 56 #c6c3c6
+0 57 #e7e3e7
+0 58 #effbff
+0 59 #8c8a8c
+0 60 #c6c3c6
+0 61 #a5a6a5
+0 62 #9c9a9c
+0 63 #e7e3e7
+0 64 #8c8a8c
+0 65 #a5a6a5
+0 66 #5a555a
+0 67 #8c8a8c
+0 68 #c6c3c6
+0 69 #bdbebd
+0 70 #dedfde
+0 71 #8c8a8c
+0 72 #8c8a8c
+0 73 #636563
+0 74 #c6c7c6
+0 75 #8c8a8c
+0 76 #c6c3c6
+0 77 #e7e3e7
+0 78 #8c8a8c
+0 79 #bdbebd
+0 80 #dedfde
+0 81 #bdbebd
+0 82 #dedfde
+0 83 #8c8a8c
+0 84 #bdb694
+0 85 #eff7ff
+0 86 #d6cb9c
+0 87 #424142
+0 88 #bdbebd
+0 89 #dedfde
+0 90 #8c8e8c
+0 91 #a5a6a5
+0 92 #bdbebd
+0 93 #bdbebd
+0 94 #dedfde
+0 95 #8c8e8c
+0 96 #8c8e8c
+0 97 #bdbebd
+0 98 #dedfde
+0 99 #bdbebd
+0 100 #dedfde
+0 101 #bdbebd
+0 102 #dedfde
+0 103 #bdb694
+0 104 #eff7ff
+0 105 #d6cb9c
+0 106 #bdb694
+0 107 #eff7ff
+0 108 #d6cb9c
+0 109 #a5a6a5
+0 110 #8c8e8c
+0 111 #bdbebd
+0 112 #dedfde
+0 113 #bdbebd
+0 114 #dedfde
+0 115 #bdbebd
+0 116 #dedfde
+0 117 #bdbebd
+0 118 #bdbebd
+0 119 #dedfde
+0 120 #bdbebd
+0 121 #8c8e8c
+0 122 #bdbebd
+0 123 #dedfde
+0 124 #bdbebd
+0 125 #dedfde
+0 126 #b5bab5
+0 127 #525152
+0 128 #e7e3e7
+0 129 #00004a
+0 130 #b5bab5
+0 131 #7b797b
+0 132 #313431
+0 133 #b5bab5
+0 134 #e7e3e7
+0 135 #c6b28c
+0 136 #b5bab5
+0 137 #8c8a8c
+0 138 #8c8a8c
+0 139 #b5bab5
+0 140 #b5bab5
+0 141 #bdbebd
+0 142 #dedfde
+0 143 #dedfe7
+0 144 #8c9294
+0 145 #d6dbd6
+0 146 #9c9eb5
+0 147 #9c0000
+0 148 #e7ebe7
+0 149 #dedfde
+0 150 #84aaff
+0 151 #7371ff
+0 152 #00007b
+0 153 #ff8200
+0 154 #007d00
+0 155 #0000b5
+0 156 #ff8200
+0 157 #0000b5
+0 158 #ff8200
+0 159 #0000b5
+6 225 225 7275 6450
+6 5475 4500 5775 6075
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 5775 4500 5475 4500 5475 6075 5775 6075 5775 4500
+4 0 0 50 -1 0 12 1.5708 4 180 1350 5700 6000 Sample (Labeled)\001
+-6
+6 1050 900 3300 1200
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 3300 1200 3300 900 1050 900 1050 1200 3300 1200
+4 0 0 50 -1 0 12 0.0000 4 180 2040 1200 1125 Subsample (Class sample)\001
+-6
+6 4200 900 6450 1200
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 6450 1200 6450 900 4200 900 4200 1200 6450 1200
+4 0 0 50 -1 0 12 0.0000 4 180 2040 4350 1125 Subsample (Class sample)\001
+-6
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 1800 1200 1800 1800 1800 2250
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 3000 1200 3000 2850
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 6375 1200 6375 1725 6375 2850
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 4800 1200 4800 2250
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 2700 2550 2700 2250 825 2250 825 2550 2700 2550
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 3150 3150 3150 2850 1275 2850 1275 3150 3150 3150
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 2475 2550 2475 2850
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 6225 2550 6225 2250 4350 2250 4350 2550 6225 2550
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 6675 3150 6675 2850 4800 2850 4800 3150 6675 3150
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 6000 2550 6000 2850
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 4575 2550 4575 3750
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 1050 2550 1050 3750
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 2175 3150 2175 3750
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 5325 3150 5325 3750
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 1950 4050 1950 4350 3150 4725
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 5475 4050 5475 4350 4425 4725
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 3150 4050 3150 3750 825 3750 825 4050 3150 4050
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 6675 4050 6675 3750 4350 3750 4350 4050 6675 4050
+2 4 1 3 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 3225 3300 600 3300 600 2025 3225 2025 3225 3300
+# Parameter Estimation
+2 4 1 3 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 6825 3300 4200 3300 4200 2025 6825 2025 6825 3300
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 3000 5550 3000 5025
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 2025 4875 2550 4875
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 4950 4875 5475 4875
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 2025 4500 1725 4500 1725 6075 2025 6075 2025 4500
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 2550 5025 2550 4725 5100 4725 5100 5025 2550 5025
+2 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 5
+ 1 1 1.00 60.00 120.00
+ 1050 1050 225 1050 225 6450 3000 6450 3000 5850
+2 1 0 1 0 7 50 0 -1 0.000 0 0 -1 1 0 5
+ 1 1 1.00 60.00 120.00
+ 6450 1050 7275 1050 7275 6450 4500 6450 4500 5850
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 5175 5850 5175 5550 2550 5550 2550 5850 5175 5850
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 3525 525 2700 900
+2 1 0 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 3975 525 4871 896
+2 4 0 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 4575 525 4575 225 2925 225 2925 525 4575 525
+4 0 0 50 -1 0 12 0.0000 4 135 1635 1425 3075 CovarianceCalculator\001
+4 0 0 50 -1 0 12 0.0000 4 135 1215 1200 2475 MeanCalculator\001
+4 0 0 50 -1 0 12 0.0000 4 135 1380 2250 3525 Covariance matrix\001
+4 0 0 50 -1 0 12 0.0000 4 135 435 1500 2775 Mean\001
+4 0 0 50 -1 0 12 0.0000 4 135 1635 4950 3075 CovarianceCalculator\001
+4 0 0 50 -1 0 12 0.0000 4 135 1215 4725 2475 MeanCalculator\001
+4 0 0 50 -1 0 12 0.0000 4 135 435 5025 2775 Mean\001
+4 0 0 50 -1 0 12 0.0000 4 135 1050 3300 1800 Measurement\001
+4 0 0 50 -1 0 12 0.0000 4 105 570 3300 2025 vectors\001
+4 0 0 50 -1 0 12 0.0000 4 135 1380 5400 3525 Covariance matrix\001
+4 0 0 50 -1 0 12 0.0000 4 180 1980 975 3975 GaussianDensityFunction\001
+4 0 0 50 -1 0 12 0.0000 4 180 1980 4500 3975 GaussianDensityFunction\001
+4 0 0 50 -1 0 12 0.0000 4 180 1410 3150 4350 Probability density\001
+4 0 0 58 -1 2 14 1.5708 4 195 2115 450 3750 (Parameter estimation)\001
+4 0 0 50 0 0 12 0.0000 4 180 915 300 750 Sample size\001
+4 0 0 50 0 0 12 0.0000 4 180 915 5925 750 Sample size\001
+4 0 0 50 -1 0 12 0.0000 4 180 1980 3150 5475 GaussianDensityFunction\001
+4 0 0 50 -1 0 12 0.0000 4 180 1290 3150 5250 Index of winning \001
+4 0 0 50 -1 0 13 0.0000 4 180 1365 3075 4950 SampleClassifier\001
+4 0 0 50 -1 0 12 1.5708 4 180 1095 1950 5850 Sample (Test)\001
+4 0 0 50 -1 0 13 0.0000 4 135 2400 2625 5775 MaximumRatioDecisionRule\001
+4 0 0 50 0 0 12 0.0000 4 180 915 3300 6150 Sample size\001
+4 0 0 58 -1 2 14 1.5708 4 195 2115 7125 3750 (Parameter estimation)\001
+4 0 0 50 -1 0 12 0.0000 4 180 1365 3075 450 Sample (Training)\001
+-6
diff --git a/SoftwareGuide/Art/DudaClassifier.fig b/SoftwareGuide/Art/DudaClassifier.fig
new file mode 100755
index 0000000000..62fe78269e
--- /dev/null
+++ b/SoftwareGuide/Art/DudaClassifier.fig
@@ -0,0 +1,575 @@
+#FIG 3.2
+Landscape
+Center
+Inches
+Letter
+100.00
+Single
+-2
+1200 2
+0 32 #f7f3f7
+0 33 #cfcfcf
+0 34 #cfcfcf
+0 35 #696d69
+0 36 #df9e8e
+0 37 #efefdf
+0 38 #bec3be
+0 39 #dfcbae
+0 40 #efe718
+0 41 #8e7dbe
+0 42 #d7d7d7
+0 43 #8e8ea6
+0 44 #494949
+0 45 #595959
+0 46 #616161
+0 47 #8e8e8e
+0 48 #b69a71
+0 49 #4192ff
+0 50 #be7138
+0 51 #d77500
+0 52 #d7ba00
+0 53 #006500
+0 54 #596938
+0 55 #cfd3cf
+0 56 #aeaaae
+0 57 #8e8ea6
+0 58 #efba59
+0 59 #8e9a69
+0 60 #494949
+0 61 #616561
+0 62 #d7d7d7
+0 63 #8e8ea6
+0 64 #595959
+0 65 #616161
+0 66 #b6e7ff
+0 67 #86c3ef
+0 68 #bebebe
+0 69 #cf9651
+0 70 #9ed3ff
+0 71 #d7d7d7
+0 72 #8e8ea6
+0 73 #8e6969
+0 74 #595959
+0 75 #616161
+0 76 #8e9e69
+0 77 #f76900
+0 78 #596938
+0 79 #8e9e69
+0 80 #494949
+0 81 #d7d7d7
+0 82 #8e8ea6
+0 83 #8e6969
+0 84 #595959
+0 85 #616161
+0 86 #8e9e69
+0 87 #f76900
+0 88 #8e9e79
+0 89 #494949
+0 90 #184918
+0 91 #596938
+0 92 #d7d7d7
+0 93 #aeaeae
+0 94 #8e8ea6
+0 95 #f7be59
+0 96 #8e9e69
+0 97 #494949
+0 98 #616161
+0 99 #61699e
+0 100 #8e6969
+0 101 #f7f7f7
+0 102 #df0000
+0 103 #595959
+0 104 #8e9e69
+0 105 #494949
+0 106 #616161
+0 107 #596938
+0 108 #d7d7d7
+0 109 #aeaeae
+0 110 #8e8ea6
+0 111 #f7be59
+0 112 #8e9e69
+0 113 #494949
+0 114 #616161
+0 115 #596938
+0 116 #d7d7d7
+0 117 #aeaeae
+0 118 #8e8ea6
+0 119 #f7be59
+0 120 #8e9e69
+0 121 #494949
+0 122 #616161
+0 123 #61699e
+0 124 #516928
+0 125 #616161
+0 126 #596938
+0 127 #969696
+0 128 #616161
+0 129 #006100
+0 130 #8e8ea6
+0 131 #8e8ea6
+0 132 #8e8ea6
+0 133 #006149
+0 134 #798649
+0 135 #e7be79
+0 136 #8e9e79
+0 137 #a6b6c7
+0 138 #696996
+0 139 #866969
+0 140 #519e49
+0 141 #d7e7e7
+0 142 #516161
+0 143 #186949
+0 144 #9ea6b6
+0 145 #ff9600
+0 146 #ff9600
+0 147 #8e6969
+0 148 #616161
+0 149 #d7d7d7
+0 150 #aeaeae
+0 151 #8e8ea6
+0 152 #494949
+0 153 #616161
+0 154 #8e6969
+0 155 #595959
+0 156 #494949
+0 157 #616161
+0 158 #aeaeae
+0 159 #616161
+0 160 #8e9e69
+0 161 #494949
+0 162 #596938
+0 163 #aeaeae
+0 164 #494949
+0 165 #006500
+0 166 #596938
+0 167 #cfd3cf
+0 168 #aeaaae
+0 169 #8e8ea6
+0 170 #efba59
+0 171 #8e9a69
+0 172 #494949
+0 173 #616561
+0 174 #d7d7d7
+0 175 #8e8ea6
+0 176 #595959
+0 177 #616161
+0 178 #d7d7d7
+0 179 #596938
+0 180 #8e8ea6
+0 181 #8e6969
+0 182 #595959
+0 183 #616161
+0 184 #8e9e69
+0 185 #494949
+0 186 #d7d7d7
+0 187 #8e8ea6
+0 188 #8e6969
+0 189 #595959
+0 190 #616161
+0 191 #8e9e79
+0 192 #494949
+0 193 #006100
+0 194 #596938
+0 195 #dfdfdf
+0 196 #aeaeae
+0 197 #8e8ea6
+0 198 #f7be59
+0 199 #8e9e69
+0 200 #494949
+0 201 #616161
+0 202 #61699e
+0 203 #d7d7d7
+0 204 #8e6969
+0 205 #595959
+0 206 #006500
+0 207 #596938
+0 208 #cfd3cf
+0 209 #aeaaae
+0 210 #8e8ea6
+0 211 #efba59
+0 212 #8e9a69
+0 213 #494949
+0 214 #616561
+0 215 #596938
+0 216 #cfd3cf
+0 217 #aeaaae
+0 218 #8e8ea6
+0 219 #efba59
+0 220 #8e9a69
+0 221 #494949
+0 222 #616561
+0 223 #596938
+0 224 #cfd3cf
+0 225 #aeaaae
+0 226 #8e8ea6
+0 227 #efba59
+0 228 #8e9a69
+0 229 #494949
+0 230 #616561
+0 231 #596938
+0 232 #cfd3cf
+0 233 #aeaaae
+0 234 #8e8ea6
+0 235 #efba59
+0 236 #8e9a69
+0 237 #494949
+0 238 #616561
+0 239 #596938
+0 240 #cfd3cf
+0 241 #aeaaae
+0 242 #8e8ea6
+0 243 #efba59
+0 244 #8e9a69
+0 245 #494949
+0 246 #616561
+0 247 #d7d7d7
+0 248 #8e8ea6
+0 249 #595959
+0 250 #616161
+0 251 #006500
+0 252 #596938
+0 253 #cfd3cf
+0 254 #aeaaae
+0 255 #8e8ea6
+0 256 #efba59
+0 257 #8e9a69
+0 258 #494949
+0 259 #616561
+0 260 #596938
+0 261 #cfd3cf
+0 262 #aeaaae
+0 263 #8e8ea6
+0 264 #efba59
+0 265 #8e9a69
+0 266 #494949
+0 267 #616561
+0 268 #596938
+0 269 #cfd3cf
+0 270 #aeaaae
+0 271 #8e8ea6
+0 272 #efba59
+0 273 #8e9a69
+0 274 #494949
+0 275 #616561
+0 276 #596938
+0 277 #cfd3cf
+0 278 #aeaaae
+0 279 #8e8ea6
+0 280 #efba59
+0 281 #8e9a69
+0 282 #494949
+0 283 #616561
+0 284 #d7d7d7
+0 285 #8e8ea6
+0 286 #595959
+0 287 #616161
+0 288 #efefcf
+0 289 #f7ae59
+0 290 #96cf9e
+0 291 #aeaaae
+0 292 #b61479
+0 293 #efefef
+0 294 #868686
+0 295 #006100
+0 296 #596938
+0 297 #d7d7d7
+0 298 #aeaeae
+0 299 #8e8ea6
+0 300 #f7be59
+0 301 #8e9e69
+0 302 #494949
+0 303 #616161
+0 304 #61699e
+0 305 #797979
+0 306 #005900
+0 307 #e77171
+0 308 #df0000
+0 309 #ffcb30
+0 310 #287949
+0 311 #df2820
+0 312 #2059c7
+0 313 #f7fbf7
+0 314 #efefef
+0 315 #e7e7e7
+0 316 #006149
+0 317 #798649
+0 318 #d7d7d7
+0 319 #aeaeae
+0 320 #617179
+0 321 #e7be79
+0 322 #8e9e79
+0 323 #494949
+0 324 #a6b6c7
+0 325 #696996
+0 326 #866969
+0 327 #519e49
+0 328 #d7e7e7
+0 329 #516161
+0 330 #186949
+0 331 #9ea6b6
+0 332 #208659
+0 333 #006100
+0 334 #596938
+0 335 #8e8ea6
+0 336 #f7be59
+0 337 #8e9e69
+0 338 #616161
+0 339 #61699e
+0 340 #616161
+0 341 #596938
+0 342 #d7d7d7
+0 343 #aeaeae
+0 344 #8e8ea6
+0 345 #f7be59
+0 346 #8e9e69
+0 347 #494949
+0 348 #616161
+0 349 #8e6969
+0 350 #595959
+0 351 #9e0000
+0 352 #9e0000
+0 353 #8e8e8e
+0 354 #414141
+0 355 #8e8e8e
+0 356 #414141
+0 357 #8e8e8e
+0 358 #414141
+0 359 #8e8e8e
+0 360 #414141
+0 361 #8e8e8e
+0 362 #414141
+0 363 #8e8e8e
+0 364 #414141
+0 365 #8e8e8e
+0 366 #414141
+0 367 #8e8e8e
+0 368 #414141
+0 369 #8e8e8e
+0 370 #414141
+0 371 #8e8e8e
+0 372 #414141
+0 373 #8e8e8e
+0 374 #414141
+0 375 #8e8e8e
+0 376 #414141
+0 377 #8e8e8e
+0 378 #414141
+0 379 #8e8e8e
+0 380 #414141
+0 381 #8e8e8e
+0 382 #414141
+0 383 #8e8e8e
+0 384 #414141
+0 385 #8e8e8e
+0 386 #414141
+0 387 #8e8e8e
+0 388 #414141
+0 389 #df9e8e
+0 390 #efefdf
+0 391 #bec3be
+0 392 #dfcbae
+0 393 #dfe3df
+0 394 #cfd3cf
+0 395 #efefef
+0 396 #d77918
+0 397 #efe718
+0 398 #8e7dbe
+0 399 #d7d7d7
+0 400 #8e8ea6
+0 401 #494949
+0 402 #8e6969
+0 403 #595959
+0 404 #616161
+0 405 #b69a71
+0 406 #4192ff
+0 407 #be7138
+0 408 #d77500
+0 409 #d7ba00
+0 410 #006500
+0 411 #596938
+0 412 #cfd3cf
+0 413 #aeaaae
+0 414 #8e8ea6
+0 415 #efba59
+0 416 #8e9a69
+0 417 #494949
+0 418 #616561
+0 419 #d7d7d7
+0 420 #8e8ea6
+0 421 #595959
+0 422 #616161
+0 423 #b6e7ff
+0 424 #86c3ef
+0 425 #bebebe
+0 426 #cf9651
+0 427 #9ed3ff
+0 428 #d7d7d7
+0 429 #8e8ea6
+0 430 #8e6969
+0 431 #595959
+0 432 #616161
+0 433 #8e9e69
+0 434 #f76900
+0 435 #596938
+0 436 #8e9e69
+0 437 #494949
+0 438 #d7d7d7
+0 439 #8e8ea6
+0 440 #8e6969
+0 441 #595959
+0 442 #616161
+0 443 #8e9e69
+0 444 #f76900
+0 445 #8e9e79
+0 446 #494949
+0 447 #184918
+0 448 #596938
+0 449 #d7d7d7
+0 450 #aeaeae
+0 451 #8e8ea6
+0 452 #f7be59
+0 453 #8e9e69
+0 454 #494949
+0 455 #616161
+0 456 #61699e
+0 457 #8e6969
+0 458 #df0000
+0 459 #595959
+0 460 #8e9e69
+0 461 #494949
+0 462 #616161
+0 463 #596938
+0 464 #d7d7d7
+0 465 #aeaeae
+0 466 #8e8ea6
+0 467 #f7be59
+0 468 #8e9e69
+0 469 #494949
+0 470 #616161
+0 471 #596938
+0 472 #d7d7d7
+0 473 #aeaeae
+0 474 #8e8ea6
+0 475 #f7be59
+0 476 #8e9e69
+0 477 #494949
+0 478 #616161
+0 479 #61699e
+0 480 #516928
+0 481 #616161
+0 482 #596938
+0 483 #969696
+0 484 #616161
+0 485 #006100
+0 486 #8e8ea6
+0 487 #8e8ea6
+0 488 #8e8ea6
+0 489 #006149
+0 490 #798649
+0 491 #e7be79
+0 492 #8e9e79
+0 493 #a6b6c7
+0 494 #696996
+0 495 #866969
+0 496 #519e49
+0 497 #d7e7e7
+0 498 #516161
+0 499 #186949
+0 500 #9ea6b6
+0 501 #ff9600
+0 502 #ff9600
+0 503 #8e6969
+0 504 #006149
+0 505 #798649
+0 506 #617179
+0 507 #e7be79
+0 508 #184918
+0 509 #596938
+0 510 #8e8ea6
+0 511 #f7be59
+0 512 #8e9e69
+0 513 #494949
+0 514 #616161
+0 515 #d7d7d7
+0 516 #aeaeae
+0 517 #8e8ea6
+0 518 #494949
+0 519 #616161
+0 520 #8e6969
+0 521 #595959
+0 522 #494949
+0 523 #616161
+0 524 #aeaeae
+0 525 #616161
+0 526 #8e9e69
+0 527 #494949
+0 528 #596938
+0 529 #aeaeae
+0 530 #494949
+0 531 #d7d7d7
+0 532 #596938
+0 533 #8e8ea6
+0 534 #8e6969
+0 535 #595959
+0 536 #616161
+0 537 #8e9e69
+0 538 #494949
+0 539 #ffcb30
+0 540 #287949
+0 541 #df2820
+0 542 #2059c7
+6 1425 300 3375 675
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 3375 675 3375 300 1425 300 1425 675 3375 675
+4 0 0 50 -1 0 12 0.0000 4 180 1605 1650 525 Membership function\001
+-6
+6 1425 900 3375 1275
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 3375 1275 3375 900 1425 900 1425 1275 3375 1275
+4 0 0 50 -1 0 12 0.0000 4 180 1605 1650 1125 Membership function\001
+-6
+6 1425 2025 3375 2400
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 3375 2400 3375 2025 1425 2025 1425 2400 3375 2400
+4 0 0 50 -1 0 12 0.0000 4 180 1605 1650 2250 Membership function\001
+-6
+6 2250 1425 2400 1575
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2325 1500 30 30 2325 1500 2355 1500
+-6
+6 2250 1575 2400 1725
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2325 1650 30 30 2325 1650 2355 1650
+-6
+6 2250 1725 2400 1875
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2325 1800 30 30 2325 1800 2355 1800
+-6
+2 1 1 1 0 7 50 -1 -1 4.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 5325 1950 5325 1425
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 450 1425 975 525 1425 525
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 450 1425 975 1125 1425 1125
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 450 1425 975 2250 1425 2250
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 6075 1425 6075 1050 4650 1050 4650 1425 6075 1425
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 3375 450 3975 450 4650 1200
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 3375 1050 3975 1050 4650 1275
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 3375 2250 4050 2250 4650 1350
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 3
+ 1 1 1.00 60.00 120.00
+ 6075 1200 6450 1200 6450 750
+4 0 0 50 -1 0 12 0.0000 4 180 1425 4725 2175 A priori knowledge\001
+4 0 0 50 -1 0 12 1.5708 4 135 1575 300 2250 Measurement vector\001
+4 0 0 50 -1 0 12 0.0000 4 135 1065 4875 1275 Decision Rule\001
+4 0 0 50 -1 0 12 0.0000 4 180 1410 3750 300 Membership score\001
+4 0 0 50 -1 0 12 0.0000 4 135 840 6000 675 Class label\001
diff --git a/SoftwareGuide/Art/StatisticalClassificationFramework.fig b/SoftwareGuide/Art/StatisticalClassificationFramework.fig
new file mode 100755
index 0000000000..5a857363d6
--- /dev/null
+++ b/SoftwareGuide/Art/StatisticalClassificationFramework.fig
@@ -0,0 +1,116 @@
+#FIG 3.2
+Landscape
+Center
+Inches
+Letter
+100.00
+Single
+-2
+1200 2
+6 150 900 2250 1275
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 2250 1275 2250 900 150 900 150 1275 2250 1275
+4 0 0 50 -1 0 12 0.0000 4 180 1650 525 1125 Membership Function\001
+-6
+6 900 600 1350 750
+6 900 600 1050 750
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 975 675 30 30 975 675 1005 675
+-6
+6 1200 600 1350 750
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 1275 675 30 30 1275 675 1305 675
+-6
+6 1050 600 1200 750
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 1125 675 30 30 1125 675 1155 675
+-6
+-6
+6 2400 1050 2850 1200
+6 2400 1050 2550 1200
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2475 1125 30 30 2475 1125 2505 1125
+-6
+6 2700 1050 2850 1200
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2775 1125 30 30 2775 1125 2805 1125
+-6
+6 2550 1050 2700 1200
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2625 1125 30 30 2625 1125 2655 1125
+-6
+-6
+6 2400 225 2850 375
+6 2400 225 2550 375
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2475 300 30 30 2475 300 2505 300
+-6
+6 2700 225 2850 375
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2775 300 30 30 2775 300 2805 300
+-6
+6 2550 225 2700 375
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 2625 300 30 30 2625 300 2655 300
+-6
+-6
+6 3000 150 5100 1275
+6 3000 900 5100 1275
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 5100 1275 5100 900 3000 900 3000 1275 5100 1275
+4 0 0 50 -1 0 12 0.0000 4 180 1650 3375 1125 Membership Function\001
+-6
+6 3900 600 4050 750
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 3975 675 30 30 3975 675 4005 675
+-6
+6 4050 600 4200 750
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 4125 675 30 30 4125 675 4155 675
+-6
+6 4200 600 4350 750
+1 3 0 1 0 0 51 0 20 0.000 1 0.0000 4275 675 30 30 4275 675 4305 675
+-6
+2 4 1 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 5100 525 3000 525 3000 150 5100 150 5100 525
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 4425 525 4425 900
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 3825 525 3825 900
+4 0 0 50 -1 0 12 0.0000 4 135 1650 3375 375 Parameter Estimation\001
+-6
+6 375 1425 750 3225
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 750 1425 375 1425 375 3225 750 3225 750 1425
+4 0 0 50 -1 0 12 1.5708 4 180 1095 600 2850 Sample (Test)\001
+-6
+6 4350 1425 4725 3225
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 4725 1425 4350 1425 4350 3225 4725 3225 4725 1425
+4 0 0 50 -1 0 12 1.5708 4 180 1500 4575 3075 MembershipSample\001
+-6
+6 1500 1950 3600 2325
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 3600 2325 3600 1950 1500 1950 1500 2325 3600 2325
+4 0 0 50 -1 0 12 0.0000 4 135 735 2175 2175 Classifier\001
+-6
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 750 2100 1500 2100
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 3600 2100 4350 2100
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 1200 1275 2025 1950
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 3975 1275 3000 1950
+2 4 1 1 0 7 50 -1 -1 4.000 0 0 7 0 0 5
+ 2250 525 150 525 150 150 2250 150 2250 525
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 1425 525 1425 900
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 825 525 825 900
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 1 1 1.00 60.00 120.00
+ 2550 2700 2550 2325
+2 4 0 1 0 7 50 -1 -1 0.000 0 0 7 0 0 5
+ 3600 3075 3600 2700 1500 2700 1500 3075 3600 3075
+4 0 0 50 -1 0 12 0.0000 4 180 1500 1950 1575 Membership scores\001
+4 0 0 50 -1 0 12 0.0000 4 135 1650 525 375 Parameter Estimation\001
+4 0 0 50 -1 0 12 0.0000 4 150 855 2250 750 parameters\001
+4 0 0 50 -1 0 12 0.0000 4 135 1065 2025 2925 Decision Rule\001
diff --git a/SoftwareGuide/Art/TwoNormalDensityFunctionPlot.jpg b/SoftwareGuide/Art/TwoNormalDensityFunctionPlot.jpg
new file mode 100755
index 0000000000000000000000000000000000000000..6d61ec83dc1799be65cc1590d49cfc1bd944ba5f
GIT binary patch
literal 20283
zcmeIacT|(v*C-r&*Fl<erAikhAWbX;1SJe5gccYHH8c&qk2)a5fT2ogBb|hhp@bHS
zN()sfp@Sd^y?3PDh~>TG%v$$%|G4k>{qY{x3i6zDc6;{O?d-7MyFUW>MN>^f4RGkt
zA;3NQKfwMTK=!hR<3k&(%QjX{R>+60R+g77ye_MQA(z!{on2gA;7Geiu9vS`TDjZ7
zt&|@*IU*l=2%p{`0Ne$fJbwJd@na`XoH%jn)XCFlna-U(bLK4DMaBzET<kpDT<n~j
zm-!`xF7t}<adKXj6BfHBB_kulBP6eULt07Vx{UPx8^D>9hxLwV96H1R*sld}e4*=L
z0KSi7$B!O4eCPyy67mb+(9uJOjvPC6_SmVDhmIZr01h2Ka`YI(@rzd&nWb(UF!Aa<
zd=eGi(Zw$<EB6@b;`S6>P{?;j7v9Mta2;Y1`8qZKgrLk_J<Dfls;t02meSw0@Am^v
z(j@^<0qpI4$M`=#|1N>!{-DN>{b$zfX*{~8+wY7MU*><tl@KYWVcS=IRt(IG+?82)
z^j-evKmQ|y|Ivi+r(m=d+6upBbY^Mp==(^bouEC;K%Ci_q7Pc|zQO+J;D9k?`--;g
zo^xJ(S@8D*_s`!Bt5J5&x9*y&&`R?TW)A}|1bU^s#_vV4-1@Qn#3w8eJ0Zf$i@;W@
z`IR{&KL)<7V`751K8Q&l)?ln*vdzxZK9YzB7pY&FoCUip8&q|W*V<k0>!*6#XI=1o
zF`Bb>UuoUcTJAdR>PPSSTB`eB#-p7aJ1V^}&O~p$5Qp!mPMp^h*S-KNG>tK-qFU<Y
zW2XeXSAi(7g!x5Ac!Zs_ciZ;0(NU;>c)^{%akWh|8mtSs$-vC`E^E4{U(6};w!2cT
zx5uI}{Bjfx-W&X}V0>veU2#c1bjwq~^+7D#pJ<DGV@uUY)~UQ1G_&4o9UUDAGkyH}
z=XVMK0DedFH?topD|;rPBL`3Ue%*w|e`fzUb;a_(R#`g|2O#J|RL=te^<@ucTzywg
zYOe4IszR<pgj9KQ&NVx@JIy#aI3$(VjVFs+kIh6fA~vQN+aMSkj+h*mXEi9pby4*}
zplSUZ*Stg}Fb--udWq9O)@arCVhgi#F*#$CNH@jW9Q<G3ke&z}ATD)fmO1H%o;0c0
zAm|3{62fxPfoax5B{l{5V(v7H0-{a8(iKRLTR`y~gJ~7N^IbQSHzDzLM)N^dOB|@C
zJ6-m#XhT+|Q4#!W(^Ie>f9z(AjxAjJnpOT^X1RIPOsDK{Elvuk-IzL!6cG_|6A|wI
z!SW$K#NG+(JwnRzAoSqjA@=9*D&BwS-Un!$x%^`p!}@bRW#bQ>f8r+;7pRB@!R)$D
z=)Ih6F(=#@j1hl+&|2XvSoSnCdN5ji2!x!Tkkl+|?#5!VxNdBg;Zayxmg3CYsMqX`
z=mur$K-5$W#*W|Iu6$}Y9kIqV1tXx-j`bJt$4sF3QB&O%tzS33VLqa0v$bs(udOwG
zDNfd&`;-{gp5x{_5oKSeh&P$R02{8og*m|uH91KU-pwn-%#daNV2tl7xCC?U2kECc
zpo@xfYrgw{*iWTjo?e>q(TmJUYMPFQq3X4n+%FG#xTZ^l%9X#9SWiGfu_8`OYKm19
z8F7bhSYOLsg6L}(gQ0A=uPBa-)14m`4x+eGazQ<>A)d_SuA9u`{y`<5?-~8T2YCN5
zAj!pJC4B4*bC;bT2iIrytTwdFQo5sA;T^znpG%4Fi`?_sr*y|)-jV(_C1^pEl1ZC=
z=M-00=bdj*geIi#2iNqLlDCX7!DQZ^&l;xL-PfCFs>>;_l#DH;bRe#Yxzjj-khn}E
z)AGuhWo%wZ7e?p~)2$OM<{aGSjEFxXT<xwi<Iz>0<y?tVV~A7{&-M$P$F;T|+!shf
zg;T+<ENRYn1uf<)8SAi!u<p>XZoMBY97W+axe`8C1vEc4{qpea9A*F_fo#T_yz5Ct
zUw;|v;2Wh$gSf9S>$)#$v6oLZ_@q5&?=3IXYXBX44spc~4;#mHpX`B(ELxQHyDlxn
zK`7p<iYNa?T1WjdN|Y|uQ_|P0W|jps--Bc2$ymbnS#_1!Xgj@UJZhC9Eo!&TWa~Ke
za`0LV7ajqw%z?bhij`vSQ8uI^1xjBH$Fm3L@ZIcWCMN?y)1p23%;oJyPp0^<OVM@d
z!#JI}4emd`vc|Tp(l!7yBWRiGDQ21<$?_J3YZnY{h<KXT$<Vpnj>5>kjTFp*N)I&%
z={eL>^vz2+Agq%Kqx9MD15FG#QWyYzzvU$QJN?<Az(xlYNl`1RTK34Hgr}WjuIC$j
zeJpcPs9XbJ#S4@w1v{ss<A$8JEjGhtk6giSqfK0OA5eU?p~SEz(&{}dwp@Lehg*4q
zi)p$#sKJW6VJL@l-olpR?Rq7-4-3iPeq6LReW&-~Jv$Mp%9GeWAaLY6Er2H%R+9U)
zWy^bN>jvi*#g=)MXV#4xQ~mhDMgWJ7Sa+5Rl1Jj6*=a@^Xm5J~U6fGBd$mhaXTAYg
zDnYQzOq4vFBwt3+P5Y$1F$TL!zTbx^cpG%z;8iZ^CAvq9I3(_~s&BDVpJucyb><e}
zl%mERhoJNB`kkbQ2EAC^KEN2Lp3ho~4G#MRQIskfZ5k8#NFeZ83pRC`MT4|l%@1eY
zNxJo&<bSA^#C@SCr%ub6e)7c98n3V2{e){not(Lb{dq3wRLij{f|lYC1LHfpVs~Fu
zBIiXK+;9$;92#ytmRRe94tH^b^~eYJ&}|x)`D{JU*a^hVr*We%JAeO<kF$)U9WE71
zcux>87SdcrAh0McS8bjn5jIz~?w&g12)sBuJtJEL3((@l4P!PLpsGu}%)bK;ABS&M
z+e*n>jxvRt`lTfBrq<(@@1YQWB5`jHKp#3{SJ_{>!{@o-J6V{uTj70E%lH1Ul>f^r
z0N{wqlRf8Vt4HA@^%_G&y$wAPTvQ{kMokWL72mDzgaPjLS3l{i#!c*$Fn<MUeQh6b
zNl|B&srqbC;*Vudij0TKLTco~9Qx+NPHn$(_1<FMiGtSBwpca+fb*Pf@@2lj^7H-^
zk{5{uypQ44kj61kyYKVlBfLAsr<3}O8rwEklT~_SaOLY;oZ;G6=vmyC@g&sjalb>J
zmS+R>fm=Xf8V%`7&`0dDe)I(Zj&iP$o{0uGXH_^1)hTT9#p|$Fx#jx?W6i$qrGNgz
zpI*Nd@PaG*^J(*-yWk(IZq*WQI7*oF=Hx;srJJnd9?SAKp=IpVppB38HJ3jr*A%h&
z8ZU;c%kG?_kOXshh4$3sn^%Ls;yKh9UCk}Z?0iK;q=EkO@ZdOG%w#ff@@|UF;|`0z
zn5MzmGgn);&ue$bPZQ9?(6VNl@b3V?rxtj8T#ac%<v@sqkQ8^47&7aG5Q-I~{c`0E
z6&CwepTwsVrHpw{yQ^VLDv%>nkzS`Se<%4r;_FC+=*tMk;IfzT!0;l;7##&z`2k@{
zvOYY5l83<QpT%PvsvkCfm`WsdC^l`#`)1Jn(c$BMKF>shnzN3pt2_o*lTy^O_PT4R
z5J@`C^buIZJzE)4lf{$|FSphntV!!37l>(@GsaO({-jSo2?CGwi?v-(mFI5x<5X9O
zAm00%u*|>g{xW{{<%b1Yb8B_UTA*cPOba`*y(>*Q{JuU&YYi2yHXRz)06wjk;a#km
z=swlbh-<>+&zhZS04=Acre^8B{_wBbc$6z`1NH)Z<wgd{G1hXoOC{Z$w~P<5v%E-0
zf8)9#wRu!xj4}{r!74@AeQq~QqJ2@t@e90V4$PwSPGb{+dFYhCb`{6BxUgQQzY=|f
zEn^qg{BbHA2XCmYk^$?2v}4Tb3E|Ro^#)uX->pk{M`M6oiY$N8@9LzLIZfD8Q%$k+
zE&2UmJN@J5gH%Hu@0<1iU@5o{m_GavZ{)*+z`(LYs=9%}bA5&rKnNZo5t?+;PS#z?
z(({Jig~xHi@*VdE#Nq`t%qGe@6#Z_9;3)l5JvqO&JlwnHo;E4$&vx#{cWftR{Dz9z
zM8mKRcLa69Cdw`vL2l|B>d`E=B)=@V{l+Ik7*k6)US~tPqCNB;7q<^UXl=0m{$-Mz
zRlMd7DlC?Zbj))|EPN4|gCz|9u^Wd2!jf`^gh>cjw}g`wuCY+D2yHZTQ^+UPFFxk5
zQvl~q-toDsNhBq-$h&<&Z_JJSVRb$z+5409x#ovc)JsEMg=Z`W&0?{cSjtlBbn@Mz
zFX+CEpY=K5W_1xhr`RjlYsPxf<<867m67ly@zVChG}Dv}6hrouCBm+vJawka-rYk;
z7JlN`%rE@X%I^_rZbL}mJVLoVRbmF6P3_4c`wYhkbM!|GcP51<Ex)&G7m*CPC}+QS
z+{5~#<F*^|jQhze3FwUB8tJAXIm;P?lHGHjUr7Q0e6?iUMk2iuUxdoyhH2BPT9A9&
zLTASM`qu@W=EATug}ke28OUlke}p{8`u*SN+kQcM>-6V>_JZeRANb-3=OFU;DE=UL
zf3p0UH(#KhRMbzRfRnG>Sc%EJwc4|%YtLT=f^&W>3LK%&9y(#;*JvXdqTenu;My*j
zvYffgYxhaLTx~b<i{c)PpS9=-+>TPF7=3kAI-a%D2c9or8JG)CtgqzAjlAyVVMbYS
zn^IgXW1&4pmG&&~rN@*bMK9)?@L_UYAkZqO1Z??|0XCIaBIMS;DbelzA)A`(EFb-%
zlcI}2Q}8&WfHhgrGHW@b>K9@dit~!PWt|3QfG|XX?BF?YOr#J$N?W8v8?70W3qq#L
z{^q`_CsiOd%<^C1&248=@#st^i*zJVi8Hs2;+st#puVj50`KzR#`(Un*{qYx%pWyD
zi{w_Tu$_wP;UCugA0hHDbF;C6<C?x<k*l(~PRngLbUuuKNEIQPz|kLC&MSybI;o$g
zmM5F&Cj74Vw+L9bTs;<B1%%MYCoDGW_N$tm-{#am=<Y~=%U<b3Jd+lmD<<$fFl-7>
zYNRx^VCyBDP?psnwOPI%<!1%{6Lqa&W@o|*1yr?<+NGsx%WO8k`9lBgKGpk|w&fw8
zW|MNnc{kxB<hW)IK6Bv%g}U3{tt{(rk`IK*%FkBRI^Qo77<SQ{kfOhQV|bsKW27A4
zE2=t>T%MY`mU8P0IQsYp$AEu}9*%Oo0#8jmz%I?t6o}bxm#&6jJmaFEegy|w^9y|A
z{GeuAaDi*2+OC+ykVlLOpAwqz;y~nlbD$i=EUj(aeJ-az21j^mWQxxB4M$48%+Sr1
zvAeBH^0@Mp9@YAuL)eCFT$%kslnQ5eWjMS)+&&5f($e|$d4YtBtiOn~JWN7Cy&Xx=
zOa~j}L+f+Z6z5oPeOXT*7i9?}?#|O&`vCE}M+Z;XUZ~Wc@A5k)S#u@p-~s*PK|jYb
zT-JNHl+nB9ivE*oB0VnmGR12PFN8KUgx2adsFoMxo(*L7hPp8T&Dkg2zid2o!fv;v
z#R<X+9wBigSC>Z_rVv6dYAyVgeE5L<XR#Yj9ixwT#R5CbW0W2^jWPIlWm7*(A50!T
zC;4bDS+ig<e1~BOr6g%HnZAvA=|>y+0+~L34g&$l;TM*k5qWOkUnJYTIaoqJ^e&A<
z@eYFGH;V6zA}x0ZPlkh5^8PpwFM!(gYlLV>Swy^Rc$Z{gQ_?-Ee$NirFR5Bk+a`M;
zN#?IjbYaQ5m9r18*O#E@0UhMw^E@UA>Ivk$Q1VBvlhAj)0~Ivw5<X0|l+wwsN8xrq
z57P2lnf}O26U}A9{J2)D(gc6iodb@(0Qjk{a$;q6-dquCHxc?Ero~CL{8B@&rkLSn
zJUK1a$!OLft=kc<kHqppy^JvkyLIqxVsdMD2?7+e;G5kT6I_SkN4JK9R7VnK)kr}h
zA6Z`fOZ<9yY<HlT%|1!0GgL%sGcl)$T&3@;v3~zRM2Aisn}1O6a;bv7`4p0AcaV&&
z-Zqi?awxc29ltr6bDl~6o1@7QS58PDIF%7^SS%*=)QH(KU@I}>KyQxj%q#bqo|_hP
zpDnyIUc1oepEOeuPIunJJeLQHSz>SFhvr|giIl8EbeQ#Mpi!#gEjper<L?O(!+Mrf
z=-b)30k4peaO}JF`THX3%@SjAQHWJs41=rc*oc&gawNIc|K1~#(ep|n-l?&g;(>y5
z_Uu&9F!60mk6(5Ixr*OCG(hXbig+pxl|d}yU<)q3xoqm{oHsapAB`#Jcd#J7aCO!K
zB1i_4<z;TN+t)Q8_`iqX#WRicKl)grn#TD!H{^2>^zwr-MJ<(}2|V9yBqC$lJwP%h
zC1QvIF}V<J7Oah|H^nExi+GWeEIdZF6Wq~O7bF$9y~clK3~*#zgN?~4Qk~DR7AoKT
zAvj3{)OU4gCpl;kKZMy%aaYxE*ReuYGWgp=@Y@1!4t52;++gXka1uIh8g4A@Ig};J
zPG~F>@k^bg&t9H0FPB<-Y*&Xg7|YEWjez*(&`^;=(O)(C@cC9Dq>6Z`*(vbM5WD}p
zc}I!%MbY7|tpgYK1^C~6`O&0Q!-sFF;sa$$Os_K(dL48c1apTdu<~_9h^jrB!G|&o
z)A6)pwLv=hcW6b#i6@rkFp-jmnAP5GsBR8OOuwrqd4>BJmoLJgTCNmXK#PhEl2X|B
zAY4>9JV+^OEYfxk4!e$SMddI0SlOJrl*r5q>UpQrSK(`z%kMfGslmZnD4N7C%0e(Q
zn{dk4k+ELRn@^N;ZHLG}(!_0r^o?GJgTR4446fTP-KVdP4$Ng+>JU*>yGIDy<n_Dy
z2lAqu0&#*%O5t)l$~BIU?`jFCT}J>3oKE3mp5GRzC*~r6*(iiya4?4V%IAX}zpR(b
z?E{3wOhOVme|3>m3+ccyqNlsIbg*=TAFigwD1ahw4tKejCY=uZkoAFX$d2vJ54?)U
zig`BO=Zo5iQgp*Dr&$D`_hb$tHo9k0DJ>q(9b%y7l#}8%(29N`H2d+rg`w~9%1_SX
z&~@9nYMw$i`~0_D&wDat@ImHYyu{G#9?v)}rd9`s&BkT%d9nE>Gs7GlAO1Ox+<4Y!
z5?2%F|AOL~IsMpRx<W<{@Q#{Iv+Q?BjFTdPyrVv{x4T~#>u;}Z)5@m0$lFq2RS__#
z&(qhLeZ9zQzl}tQt@#zXtU;Bzsn*-EGA?KaPO9~^NS0>mzXqwlI8AvtED_x((U&7<
zI^S&Ug{oBIR0h9gRh#3)IfF8WR=W1;>FZya4NGh}>dK8RfT#O0Q`{g9P6u9Q&?CB>
zF7MlY5pLH7VqmPKJaT?sIUQ2_b?uAkJ|7*vYi*DJGX`$03a%IVTy^G09bp3dfUc>p
zQ5k*z-@~f&oP$uCH`Isq8ja?DMkdA}n4FaXR?0W*b!4yY^+y8^?xW2h-IH?-ywofj
zmk2(?&kY2Xz@kObZ8MY{?EK=;?}z>+wzxdI4=7(|%T3KXg~!{*Z9MY8c(ob%W{eyt
z%;9s&c8-~91$k(yXaiEFGgHluDTPM;LZYDKxh<&#W$!^w2WbI^Q`@?v%CC;G{lGEm
zHQcRN_)D6F5Xc^k*#|uNS~fk>)ar?+V2WCYL#-9$h~E7<sab>f`%EsrFLYyj2<f)x
z!-8=g1*$YuGtFRzCieRQ6+ed@hfdhLTZDK0hVRPPjPC9(YS&M0Ifu3<uTO>}up%X)
zKu^zKLJ9_C2mL#W*OX(Q*-Y2nrdZm;E>^XKu8SzHA}VKe!WR;=C!CkS$Ep7ozML1U
zzg?OL9DJ5g5%Z@nsk$f|lPZLQK962`Gd2|Z=GQd75?T;TF?4TWmD09Mf-FbhI-vBV
z;@$cCy{6blnj04@aNA{_QJrmK_ty`^dEk<Lxh1uXx6dH;=`J}mYyx)b+d=u|vDv5o
z7nJishEzh`$U;J6B_q}K9_{lF2l{tc!MSx|Z@_+!?=$N@peEq{hpLrrvkPpI4JHGK
zirh~4fCqPN_l>;!oF0Qo0qwlCxJ(T1{7iSY<1N6ygw)T9lBW|GL&3sMWBf5H9v7yd
z<UD)!YNfAQ3OKbjh@|}`mkk^sB%hF$+pAe!fKg$rWWR?6r()lAkoi1VYQ@nlyU#@)
zJ2!=4z0Qx&%|Zltu78#|9IGF$z(80pqFL&1d5mete7DE9TYVyw|7a9Co=ZrR-eeQu
zE!hWHKF;db{T?s;on1EN3WE8S=kC&jeJ4jsj<$d@GzX_6^wm<U%P5Mq>WT&`&(^g2
zVAc*WChHG3TS_ENV(59Wc!d*~tJdWeu6)}?Inl!er^C>kDpAgxuy~t9bVC^D>neCS
z8ijTW7hk<JWEJ0z96&+k8UOHG#$tJe%reBxG&!r3<1RbAfG4iYM+%0+dui}~>S7uu
zyZTD+zvcM1@$+*;lF9?m>0E5xq6Tkzq|k=$g^F)-q-RpfjrLNZFdb1-*AJ=tfR2RK
zM5TKh_qw*|r|m>-r@GlGp?ESm4N*C7maa<Ck0^N5$lYd;>^|RHo>z7qnvoaNv`g8_
zraJeD|9+ruhs`>4;`j<n?0fbBA=E9+J>gf`n?2iec0J*hEZS_kJ3Q@1fl=3~NZzc^
zNtX|{{2cZkkt;Sz1Yfvd-L1Cs$Z%f1gXm#tfXfx(6(M5ef$rH|>O#Il^QRM<^PC?d
znNx~#bgR6Jt*jc#YtnPj#OtDyo1KqaFkXz1;gR{Zp&7IeaBTN7Pup@S*@a{jJ<rKu
zm4C-JvhMlBtj@<Mv^{3LfXN;E0?koX?v`0`PK~T_M=L(Jv_B7LM2JreOQ$gt90(3c
zd$?*`dvhg$%WX#DU!w9Ih#Jw9rb<tKq1n$1w2U6vp?x6yjhfpytSyNEgM*gReL|=b
zUxfs?U54i>Y%m5hP(`bvljYfR^8@1IwwvG3Ghfq+Dz;a}GnHR*NpK83m-L@Z;N3bi
zbSm|$u;@!9=WUUfVn#~0Bc(#Nv-V<@mG1S-?*@MJe?J-NQ+-<mXkSJ8X0cfO$!X2c
zxoQid8EPW=GBXQBBUbCy4>p$bjQcpY5@JS!-pO}Q6<i8PM?G{gCBKS5*5^@xFp51<
zmBJu*dC$D)2G&S9zp+{lOmz%tmyM>#I>ueJ-IB%PGF&e%spC4eOj3D0wc1(~#)`cO
z9zB{XC4M>Vs^~owHx&g5d5g_e!<OVg^s=WbS#;Hjwx(G}7!;|=xD1H-#dK8X&V%_@
zx`~`YT%29hlH<Ey6~r7;?$p?o=kXcy?GT>yG2ttFYW%p;ME3S<8he%|GEc3nam6^#
zqfcC&<4Ooxt;^B@0mOs<P@TN{B}k)>!`$8ZbA4M*wG`LhiGmdd1z4lR3MN_ZAf*gf
zyU$^XB9ALGv?5(aB=68u%G(J;1I}g=ukvG9Jp9C(a&2PR8w0A6P2L<p0~|h!8{}|4
z@|m}>%P_IJGSDM|W^hxR@0agK>hCvjypEs>qAw`;QE7ARg7#5-5Mutx#Zz*w@T4a4
zkRGHbsPYWKAivlw&+P0N%L9L>Se4Sms7yWnX5S3dxD!q$W_pODIxiu+JnLv@_y@Ke
z*=OQ5z{rMAk7XvD6x7sZikHtEE^9*-c;FqJAzIvtP*n=+n_v4{BCQs4%fn+J%Pl?p
z*gAu<f=X<c0WTjWl%hs3(P5@%_L`ah%+W3nbg%crXI$&T^R^>}o@xL?C6||JBVQRm
z^mC*UmBh`MkdDlK)o2cK`$Wn$Z^v$S15v&i9{`sfO3Gi8t$HGE4Emz;s+Z$GUtL`i
zH4NWId@pbQQcPswypD7PeXMJ1F3*~ka*bOE=@5!84Gj&>E>B$*;ra6`+P|A!M%j>v
zIBV(rI{PwyKY7dmXMifUPBq}Xl7gq_+4bhp6g`F5so9llH`Y50tpX#}DtkoALi=P;
zO`#~YN{0FNHrkC;TgO9!d(E?Zx*y%B*9@E8T~%qh`r+!6NnFdRIpC{lPX$3|9l~V=
zXAJ98`SjidC!Of~XOwbMQKqs_0xD6VODmBU*zGb;bw)!7NZ;k}?&6n+8^;HF!&1eP
zRFkUjUrghntBK^iWaJ_MU9@4HR#;5Uof|+BMZ3*E-lK-AtjM4LEc2-WKBN0YCa6Ly
zRAv#F9h@@q{MJ_~05)T<P6RGC)%qfV#_NPTsf+r24$%ZyN>F--b8kagR|aym%z}Q?
zzB8T%TMzsFUu9RgQ@!jWJV;C<!pj7xf5s%t#cYSAnGP30Msj*`b{4Ft^$y+s^7fw!
zoCwOS$n}6oX97QR;rhM)zj_cjh+K@0N0C8EmEPxUz&|bE(aN-H2|1g2!#qsi;5}Sc
zkPTK^X1TIeAwT*`odp&?(u2mGxnf-@f>PCjHoTes#@FHVT{8xioHbCeV^L$e`5?)V
z)>Dh}uAaT(3Ao*4?}L!Kw5o47!?)0u0+izeZJDkiJ;%OVI>6=bjpO|=d?wlwS3IVK
zY)D52#Qe7XEd~2IMx$f=XIbd*d6S`m`_JoBH(uhF8HGYZN^8wy7KXLce&A4Y(hu4d
zp6a#UVXU8|UvLu}=>cN;P+TDpm6F@6>pXy~)UAoj=upKrC^nNSmql0rmuEjKjH_Rd
zLr(Y2Ziw-Kj_2+J*jTK#m~TS9nF;z00FhnaTIE`B`Pf{3huuCvQ@dS*5W=v%4|w?f
zM(^*c4>+_r0i=fmI(StI1gfg4@-KC{TW^3)wCikfj~zNU1l)GJ*oppIELKaYuYJFX
z8*sTD^tzZ$B<!8o9}g-o1~`_3M1;^+%ww8q6{kqI(f*=GkW?-8HuCC|u+Ph+N>mk8
zRIkZ<bg8v=G>0B-cpC)fi}qwphoqMIBFeElS+?H*A9*)-1IZ+&``6H<nCfJ@_8HUA
zu)WjZKA*y&H_-VKXINey&dyi5o(7L=!Vha-eEr+kjX(3aQ@EQ37p@=7obTil`-tP$
zctJ{zEM(u1GVW0u*5Sy+V)9!Ju&#C=syiMNXDAId%o-V(a={h;*yafrG!*8gUe@NJ
zeGcOJ2z)~4AIJ8v^#`hD?TF)GVD|Po3uzMT>lI}kbV<_&P${cQh`YsrH}j^PXo8tw
zLq(;8KE0)doxr>We)EmTFYin*Jlr|d#qkecY|1LnXL;ZfuRPB_0Ci^MNo?npBF{v#
zg!VgmWWz;|q**QVnE7Gt)NdC;bQsTc^mvL{b5_;SffBXJY-}XX>QOSG>H9t0-wOf&
zK!9?kl@tdI<1Z(5LELdk16jT7U$^m&8iG<ax`K2{CE{URqmUV(qOft2Rsx>u45|dy
z7O5YPS4&IW;@sc`Do^tzW}|lXwG_fhTcL(M`LQW{rmGaWD?Xck(Rzx7P#jfu!fqmW
zqWq>>Bw})~Zll0Wo<ZyP19lG`xtjB`-`7lvfs0GuG%!LiH-wdjENff^eYbD`Zdu%2
z_rx^^DW0=S<?`&+xjllRDaa^U<Ce~mhe3mh8BI=Z_s`F`jKPLzTjbycG}-8~!vYCU
zbJ3g|AIK-){uIPzm<$RKp)M(Z*Liwf#ZlLTpQ+HU1~bdwL73+ZZFF1WgH^BpQuXyU
z{5=N+0N%G2UX(DNzY`lE8eu|kh^jGv>FVTT*zIh`57viu16e`fKdwH!3PRP(=1OUE
zn26<>O?oRY%Xi4Pp&`tA)ez{K&-~S}w^#pZ(=I!_Cy}ae%qxv5q@URbICJMlXfu+g
z7^nV=+M_3{6J*<L%P%+AG?z?$Qnla7i-_u+9CV!%*vkHG+zp~iQYa?EJohjlH7VCc
z-n`{aG`Pj><YUjKBGJs3-68qc&63zLf>k$9{Q8wY0N{fBNO-H)<@n&u3`G8%*>j*C
z=#E;`dy!FHR2*&&p6(``SSc~M5pK@$DmS42Tj2fkLHbWS8SIKZIN%J=p2p;K*uh<T
z3~({I_Q_ui&0_;1Gka)-E8<_S;!#AFjOVAPDQ3B4esfu{#*Uuqr<FJDJyMpXG~GmM
z)DoFR^VyCv6kqL4I%lt^s$F*Bx9`jy=BC^h8SU7$7M+7zIrd_P<DBfPYM07^KK|{a
z%?*kD<-X6hG(D%MoV^sd#K`!F`G{zkN-;&|2L!<H3*)Uscu_}{*||Ei^!$q=SuyRn
zikIJv+<#Dvwi(-V8xZE7!aom-bbjKX8JuSxlRuxe2W9_GB>>=Mn9r=)(U&yEc`e@(
zJ@cVE6@lAeyReR|#YI$V8t37x1Raw#nz88@rR15m*z1?g#WWp%Y(FGqcU3FkoDh=N
z!_7vv73NJtd0>Wo&v&}{g7pfy{Lf}e<3Bgy&b>+y1!}RX3s(4^c01Vof2RIOl+erP
zjA4Z@^9cfGKs0$c&hA6VD8%e!GPAt92al}cnX~RiCWWH8#z?o8HK+<Xv|K3U7wiwx
z&hB*t+~|G%yEDsBR7_FIdOKFMb57pyJ$ioe?8J&SoFl>cWC<x~B2<`Jj=Ne`>m!<*
z_P=$^u<fv~IKhSfK<0f?Az|&?8#QqiU0VfwRjNL}e)vyWP>I@e?^iI49y3dx+x7>7
zI*v;=QC;={OG1Bqg>eALX5GUyrg`8Fty?j)*Ul5Co;MRUQD(O5-JiS2yv+1l<5+Hc
z0mYSt&sAF$#tPE9A*q|l0rdBIm8Ukn54fR=MhuK*7RasbsS}M?=_clf2<V3VtlvHC
z#Gi<AQxnQEF$Y^kD$+<8YtN2QoiXeW>{c$2BEZotNM2$qOh@nC^8=!OzA!j$QQxg)
zBIzzH71FAYuLIW&PcFpsqw80Z4WwOmWG={kJqMeI;mt8W7d7E6ZAA4FZ+5qLT5z@}
zw``UW+Z8;3ojBNy)#!gjB}x*5^j;~S*z|idZ#mM@GoR@$Uq8MOD5kDvvXN@IK(#Wt
zt~UP)t;2vOv|0dh6s){=^Q|B31r{wgD&KoVOddCClK<_a#bka>W<hmbnkVmGC&oDF
z(#%`BPO~vNs!8F&=Tx;Ae%Zd>gMQ4R2jveV>G7g6?4KWvGyhQO4}q$WL9f5D{Q>!>
z6yUN$bINT;ht=V2?rdTV^dkjH5ZE1J0d8T|FbBl^tjrfC*i47i9Q~bg`@~~w$n_1d
zovG7RChZEcBnT*0#$cLeGvh<qI;U2*tkS5N^!|>5lyeg*h~NFwUd-;cN9la^$IMFb
zovok!`f(r8os}efAe@Ldai{Gnw6m`25ZYWI&7Z83&dXf>A($PADhL8&P=Gwxd!MBU
z%P)!)@_JYicvOFbh<5v^osM+g>9YC#!?QSnh-adxe6{)w0Y&CeJdE?>5wv#<>z78*
zyJ0rS>q8CYXPQDH)AuHdgE5Fz)%MAhf_1F261X$_Ym4p|j3+rXIHtrM|9nC)k{TE!
zUBcuNld`U2gqZbJ)l)r<vW(p2*<>EX3K5qov&YMDSb5ghzkdF*@gI7|DxfYZ*l9^E
zL-mVxC2w|z<;PohhYKPd)%oI^)v$A`-qWN(=5w-bHRxw2j(XYP=F(EIWs@6xa)bG4
ziNV1Wd~!Mv1qidYPnUy<5A3`FjUQo)fASYrv7R)G+^zSpo@q95Do&Mm(smp~3D5-9
z_$p@;73+yAZ+d>G6LhF9wy`d_Ud!E}Y&$rZ+_#Y7ori+^Iw9!=NU&o{{r&LNjX@Vn
z@&MYw6kU0SH&t*tz$zPM;fV?SjlTWsuvrgPr&>R^5tS!zqo4!mu+%yTtNP<<m)%dk
zdHNd}<5Fm~-Z6W6&IQCHY=c8S1V8_9+(QQUNAumHbE7(~nthT-;1ozCp#_N**=rue
zE&ST<nwk9_Jm7?;ZKlU@ToAdrFcO=74)Gp^Tzi`r9!5h!7{|+!kICHnvh?TR;T^TO
z-c~k2=!JiP%_Dr~<^ZF#gG+irIEwJmc?z!TlxMIw{D~q%Kjja?4|5=ECsX*NL{cjk
z)cs=`t0)0}r3=1Y)ApE<x4-=LkMi4DnvF!e4XGrH8HH}ak0ifC^k}RY)rE{A%e>KA
zmwr$#FYU#?ndq>Qi(S@{gRx`EW<H=0Ze<W11_jEmpZa!O*M7sS-{xrZ7>w*|ce}S~
zOU6_e)Q*n9XuDQofD7?_mhCr%4SVX>ai%?j*;KP=-Q$Oiq}*wJ&Dt)cq>7AFX63mQ
zVyBk(5IYv=YZ6+u`Qi8P$Pb(Ouu0rW;Ll!HN9XWr>#Do*b*@3RDM%5bFW9W|g6T~I
zv3mK~RCt;{w()g9b&y(KIK0BPdLNMU7Q9G@@^hCG;OC0|p_JBvge!Kk0UX!kEez7~
z3~jUwS+hXXK1KS@&%&yqC@k@}3|)3&s>82$YQt~zWB82e{n*@ZzEgPQHBe9P?(TTg
zb&<Urm<atp>n4*`M+ymnjG@<?=8Lg*^lS5#S8-M%_0!84zJWH&L@^Qj#>UJ<UvYhi
zeH5PASACdOBpxBk_sLNF&RP(~4Q9Zmmr=f!#iVamq;FOb{2deEdb|PYP?Wp(ae3R8
z;K6hUEl74qnvLor@rL5NA=Whp);kRah6DUg_Pz8D1iE9;@n&W^II%y)UVxuV+DBh>
z!&YMzWOyB4;G@m%Fq@qN2cl=X>SyY_dsM+9&#O2Xxp^_Qk>x%YE3FExRTn1E-lkRO
z(^V%`RaHj<e-H?;#y9O|S4v3W23tTLILWrU&Ny$^H+75t15x8yTysO5j|;92a5~uY
zS@0QmX9joC>tKsKw*ZIENQUQ|P29AA4_DRm2@Wfvb0na^c$^>{*bJVmDNb421}
zUo$sFKta5&@-i-*iOhN6EM9tBSsnBFzrN!?)PWI>f4<vZ@R!D-zfBgO5QG)G-QdX!
z*}#3qmP?)+l6ga4`hRR<IO?0}_`&3Xwd~Qd7$2R%*!34raP(Yr&13{D;i2n_^&VZE
ztCu#)jCZsY!z#39Wrp!)-Qo4eF}}uwBiOOn&hPX9RG4)b=&3L`Mh_^^xQ<na)NAnH
zupGU4I~4C0arO}boXo~pj~c5hnd-zxWSvuZBz^h=k&W|lCJG^>%RNHBtvnkpLJ?hP
z8Pzj6KZb}b<KU_VvoDSG=`gB#+?foXXBV~ov`mp+O9DdkR=CQ<w4G*ORw~d7Ff4?9
zKoNzRY5oT-N)t~-uW1}zk~XwN<S$j06I;zrhlX_oA%ul`yQtzf?{lc+z3fBcQb5)8
zhNyCx>45us&q@WCvfu-J?CbKJ4ZNbJ6biG_Rlk3ep`fK~#Ica1zbp$4r(G0N6=mZ&
z;>a}6)dqeS)jkU!bB=b-9-y1LP$1Lizqf84si$C9aT?ZxOK*?bRcu*#1M%9Yv5R+D
z<t>Y!cwC!}L>=4c6J@a7$eM>EA8##aDV_Y3C&{SCVDu2%fY@+u77K@J&@=w`$R*@H
z;4}EdqwgFaQWQ>j9jqqXf-P)*$KDP!7i%7|XAIq`Jk_23M_nCNHbiHuKEu>Q(_+kh
zxp9OxoXDBd>+CZMWfzpn&(DSH<RW;1^d@k|HwWR>(HEJg#QU7?yqHlJsZTSpLk+7F
zZ_M8t=9rxt?ZD)PhGr>os1=PnD*qYyYWqB_cVf7^Xof=K-R_Z`RRk?MyMAlsrIYu?
z6}Uaxcm#GGP%ix(ik<kB;XdonYzU+G)dUn&e%c4vgeGmata0W>JQcdhfld`Gnw5fM
zRM?5$dfOT}XpVm7KfUmYm4R+O24Y{<n9DEA-5;6aMCsSov*x}M_g-~@B3T8eMU5%#
z!^~!ddId2#U47;4A4&(@2B!oxFkJ}w_vrUXKE2^o`KI8H6xGh`g1<Y9L)_hzhlmG%
ziLYy0DcE-*dgHdysid6?$H+Fzrc#^5@KN(Uum+i+&GFD8uiCD~oMFnu+J57K2QpT$
zX$a+J8Cx_M;wx2)k;092`n;okO9|<sSWd2Ge(2gQ7=9_CV~Z1o@X`+_nS{sdl$OrY
z263qpHzjmc6tL*5n1-6y{YrT`;=zxIaclNd$}!r>)Ip0)G+|5gQj5S@O=Zf%U73`Q
zUUbuzmHC`<f0_Rt7wAA0P9!egMVxqRadkX!s(h1*kk8XToVxmQps@R~<37M+`8@8T
zPrre2Ayk`DPv7UJ8p%z?a(808$hczV58`RL%sMW|+A3xC#8fY%R8R1Zy9;wHMQV$^
z*fe<y-4r^$gP#H|3Ew(EMgKT)T>IQz333PZ<VG#2U9&-yKV+Ci85@q+Z2vVm#{D_-
zKHwBWtkQBaXnI+%gc_4mzJP88EtU#)>*issW(I1N6%1y*Rg-TTVe|I^3FZveMC`_I
z4r?}{p0EYWkjZfZIoszSUhT*cpBSy2772!J-Gb%(4WGxY8D##77v8qTda!xzsGi5Z
zT-FIz9Z!F|GEPbQ%}5Zsp;kc<Z6d{d>{%<*+Vv*(fKNbnq9!X*P6~5)YL7RGJ^N!E
z^vV$%f!MZwM}c_Z-rT+`nJ9<JU|KtW&r=c1hX-wgWZj14(k2b3ca<EaxDOaf9v<kh
z_0~rR0r2Nk=+gB|1(NEz_@FH9;9c5vJ)!d4IXP{X?i>R_0uT2})x?#}(pgJ#(okX9
zY($`Bb926bO@@IV8H$Sze{t%VRy?^Uu#%z9Oc1CPyz()Hw)eVuD|`3x`{jMWue6;m
z|I&%AfNbO*LdDc&r@Bi%0LYUkzh^$pS&fP4;oJvEuAMh;Ri6nDi2M9pay)IVL8W>3
zZrqB_=du4gP)blchqM?*YnRaAN?}f=fOVdfCVS<$Hm}lvP?hLzAIZd(NBaP#)i7I$
z2RKrpe01mfc2f0>2zSMjPoa^yJan_6HooE+Tg6apGr78;4`IJ|{q=b3pN(acl-`{I
zh5(~Z-S(R^C%VdVu_8!Cadg0Q|2pe^Kswf6MKtLn0#pcY_l61d4rC4Gd@gZMBJbTl
zfuc9;btPO~pV{p-W!{GHk$N9dI?c+g5>)qQZlr{%+huAAwEY^WGMAsHE~HjrE{gX4
z{4zjDOe@QMj>Y1g8?iwBu&bv99F`TmN+|FYS>fLqU|LC?jxvE7qFHOVnA`RN7vJBp
z{r#&_A9;QysE&pSdD#FdAuSkq8m~E%v5BnI8;n*WcR~i`3Zjc9?>XOF&muQ6w`%hq
zt)2MI{DY(Dxc0Gr9hJBH0PU!y-La^Ca?yqX%X_At!R5DDw>{GP?!Nxa$?)3IVa#I}
zB$DX1#QUat+Lt6$<N$}bBU#I!VQ@0hb#LXD-6(owmO0-%;9p&cIM3trVLM9hC0s6{
zpwEkS-9%g~px#v}p@>aRp21yFwhPh;!Ni8_1D4N+dsg=45)yGc58tzrfH5mE0WUlo
ziwr*b&lJODaJ$h@8aL9j8kRBdtyYqqcs@nYfXU4hN`oqO%4uz|nBei*HW)*I9R*X3
zZ1G#En{T!9rJ0YN|0huSPaD<+RY=0pb0xFcZfF~1@7hGz5T<FXsk)x?yB5?WOS;>O
zPKeOIbm8)U1xt#%O*SxR=;6QxcSVV*<P+UNkZp%}u05ep{l>S1#~>mxN{b<{1pLl#
z<=)ypfXmlr+IZlG8x_|ZqY&9D(p2dSv?|VemOWQPEzWaDd_=p3uF)Zr)){6IC6(qB
z4Q?EEGFcD&Q`Zpyz;>1bdx67X+VrlqINj-<ZbG#-Ygx53uw(0+x*+Wgz|OS0w}bG<
z9+>`0d0GK%H_Ml;2;<|dLt+)ZbbwIaw79HL0;?;*>tDa&_uX9n^z-Kuvo%j)8KQ4C
zQ=K`w;V!qvQlpG}eqKCzyuNo+xu7uU1^9GW>ddIfNdDQkpLqXlG|-1r(dOkSkjwA{
zNiFMm2~>pFO)yyZ`n=y~4t4ZjEy~T03`3bLUuv8d#1B)LdHA4sh4AEmzET(CeiH~k
zo@3ocf>6j?$7mv8wnc7G*((ki&TppPq&L6{^quWV%9=ZpvJ89K>ecW6LH@t>*@>rf
zklKoV`vR}LNmqLjx#+9~9UW#j+Uk_P^8XL0`@;l+^kETgDDcXNEhiR~L}w;ov%|Hw
zc05J6clH61vqAQ4o$F8DZ%L4?E9O*e=?441@!7L?#a)ACj2X`?SHN&fU(JdPnT1z0
zc+`tsFyi&X|5rpc)pwfVj7=CFV(s!(>wJcqI$`V~Si+eTT|VkBV+HTEqFY6|gU1P?
z+M$yBfR}zNd-n_22#VVMCse4UE0~p{-9KIa-#%I_B+3cb_Sy6)bsA2wU6d5o@|h^R
z)HSMNk{Ims-%rjDZZ8PUflBdEyhO7#Jn~&x{)_+!qz^)_g3BL)6t@0houyRdU4+nm
zJ297J?S-0+aAIP^u~@;CSJ{}ogfxiE$XA=h$#z<|pA?D#B4jhNv;>!z405FlpGQEm
zqVx{#NKehDiOey1oiidAr;Y0OR1HIkAKim94mJyQq=`G^fjcRo&AMviypl|6`RwIe
zU%w!J;y7<~k+8;DEX4(H7%vCaABwc8tb%c3Hro#B6^+rk+ofSd3xo8iTbuoPX6!f!
zqj|+Tj`7c&vtWIwv@FqelEQem>kFbsR#W%=qaW4c)nzGOabB)kz2op9T01W5#ln~A
z!+!nJeNAKHQid;Z{NdgrOZqO%Iax89U*^7y^JmIh)TCR01e&mZzC!9<rgb)sbc(B1
z7lh)uIO6d;eaVG0afmUjhSS_Ddjk~Ex0Cz=t_QnO=6R)VC*w_?Wa8EWN9*=>QqRx>
zYDoph=P5#%UKKrHwO6-JU(z-+NK1<qO2e!!re+;Gp!i5*_iv8H`bNgF{HP9DWlx+j
z9%PPYZryHENVpwXH6xxC_u7i0=l$-kRfJ*GyE)d%NwbtgQ>YgBb*Y=CW@<^3W9~B_
z4cwbh9#0_Ha^ZeDiHF_~t;sp6lv}9jBa3Td$RQWgbY?0lVKZod(2bcniQ_c3`R>*+
zn;i4ccAttG)LGk*5+O8&NlRNALUvez<`a{MMPVxN=+K<bI&8Kpoo#w6HeZlXvuHoL
zm*SF4YiAGXz*gSVA{3Q!MB9)|<WR{O0aIC_om`ye^5k-%=uLmhELESkS}=gA$wl6T
z>$xe4&nGH_p??w^XkA8tn>FxZs1#;aeSMW%bb`m6XvWNzL}UvPSC46wd9NmQa6j_E
zzsI4=^;79?YHe|Do(o@m4_r;TbxPHJJv0YvznsKU{OpXTal1+Lho|@Ng~1~G;0Bgv
zs9MMZafyiz7I33LKGBQ|T`d%u1uBr3;j(tG4L3jREP}{Ns@P-ygECqc``XOpYS<CM
zf_ep*r{J{6M*n6+w#%z9HhHO+R>JpRqL|Ggv6rSyvcS0@%FVIwy#QyA_x&1c8ssDk
zpY2VM3<?btewa7K!Py#VMVVLakO9TYA!w=#toU=AbSeOW#I=#a-wo$LQIX|_LF$=O
zq)tI9%3Fljqo0?^?BO$~aJJK;puco`#h}TkZfqsYqR`>3ST`xj{-z8vpIneP+aa%~
z4|Pfv&glJI(Aig-7JH)SW4?vwh>r)12?^fF$0ljMaM1(fWh8#5<Fm1@S9`x6c30>e
z3$epuuFj&9x81ZN!%2(aEyUc;lPa>N2rgvPK7yQY%@iz_t+Qt;Hg4ms$;_Kf5PzO-
z+htM4z>O(TcNWa~3{HY*S-0{UzqgW(nKa7KL=Z&_6oE-cV()E1kn{2ku4+}XGnrv!
zHTrjT>6i}%TVFigiz@wd(CU?~Nzq*I<3(abq+I6{f+1Rm9h>)YZ2MlN<rbos6x<|>
z_pM7*av7U#U{RA#QXFFbBeKS@Q7_tHUJ4_K5se`aA5XA?dy{}(qFQE^v!T5N&Lst*
zd8Ce%X)cnt#A}vbj|43cXAW))Vu}di;tAxNrmkfy{G#|9m<a{vBd!RMrI~it=en=@
zL=~lio2Y`}G}1zTL~M&_NOGb4NiqJ4M8}xmDX3~(1T#cmi}5M9&YPI^Aivc7z-n3P
zt~jG}?jO({S)yD9(~~70uOeI-A0vgTwOqRf9Vm*I=3;8vOUwK5nwP9$Zj?C-zS5oS
zj9YXm-02^HH3T}$)i=M4iRCp+vy|3ES8`(Rjn{`GiB;QACGQBmsP|w2u4Jlp%sG?@
znw<B;Tp9qf7b;)Tl9xf;3n4Mxqqyh7D8mq#_Jh{5>0NV$e5hpKDsvlq9m|(cBR1yZ
zT<q=yk(p1`HP;=Ro%Ct4JeB95$|k+7_wcWld;y2(>%G{`PmRB(7ALVU2<GE-!rNJ&
zPG^VIWH6t_&njnMTh{ewaUUShwmz^jfIe9+zz1)^j3+WlC`2O%x=p|5u_yMmHtQV&
zn@K&Vn=LLKEmX<gg6Vk&?I0(ZAquVMHeHcx$^g-eg1NST)pI)Q|4#?~9yKZ_ICZ8Z
z-EhSIRh^6BMBd8a3k08WC>A@x`v_bBJ6H6Unp<<sYmRR`^4_aLon)3=bXv?TR5`c{
z$;X3(xGg|`4|BZcqF|t|9C!0Zq^pHaink$!+~sRmmW!vfY8h2_Br{IQ)|@Sh?YQr#
zBeo&LH+X${CvQTyj69FT%zF%P3@=S8WcG8yJSDWmlX|B*9JeTN53eKW>4;9SO1sL_
z$D{3a7edyfn9c0}YE+`P=hlN!Pt~0&$rO7=wT(7?A6I)fGv%XO)&0j?-$J_i%O0|~
zK`d<^wH2tM`B*?ca&oUw8kcYtA#cf*C-hBVL9vS38naS}bbCI5EE#8gG~^QA{#CgD
zGLcO0GQ{%mG%H<fcPuJY$nTEDs78;|%3Lj~qYO*r+6S_6nG5PozK{WS)|FQ+em7?C
zbA9f%c9@~NA+9t14N~OX1t`Q@VmWh~$e;%;YX48(UP^quOXZ~*<X|kVyxXJ1gGjZu
zB$XqQ6fyRcybAAA_X`*rb(8GyxKGankSpzcozeGCJ!DQ&l9^BO>nZzySqw4XjhDus
z!u1q};f46g7Ri}cUMkNyyzpLbv)@P`R&y$2ALmU=s)5O@TfZ?1<8e>-cx)1zmr)ne
zn_*};4?-^@G{Zp)4D|1fOJK`qa^_z4hOKA62XpR?iI0A&GmZ7G$So+AL0?pL|IBAv
zfm@?+5e&j=XPahC*${oUnHX05;2+R2hqXw<-f<9<ZBIT_gS{ehrhTD21}G%7Eswft
z@+8)l;+E^zudR8b8ybRpctwi}hY@_ptMYA-;SF+duulKEuxBbxw2;g|16!>0jjY@#
zwVvDj??FBgS07ljh^VKl>vV;Rln)0tP^zh+J&+=Op<-R9p`b_`Lf}`-rvGV{0F6pq
zOmhtwCA0RrU<JP1l~?B#?r?#en_qs6GDlR<PJ{YX8kFI+Agua5Pwi{&{U>kdf7Sp1
IJo|nB3!L*JTmS$7
literal 0
HcmV?d00001
diff --git a/SoftwareGuide/Examples/CMakeLists.txt b/SoftwareGuide/Examples/CMakeLists.txt
index ce2f041bdd..8af106d77f 100644
--- a/SoftwareGuide/Examples/CMakeLists.txt
+++ b/SoftwareGuide/Examples/CMakeLists.txt
@@ -207,13 +207,20 @@ SET( OTB_EXAMPLES_SRCS
${OTB_SOURCE_DIR}/Examples/Patented/FuzzyConnectednessImageFilter.cxx
${OTB_SOURCE_DIR}/Examples/ChangeDetection/ChangeDetectionFrameworkExample.cxx
${OTB_SOURCE_DIR}/Examples/ChangeDetection/DiffChDet.cxx
- ${OTB_SOURCE_DIR}/Examples/ChangeDetection/RatioChDet.cxx
+ ${OTB_SOURCE_DIR}/Examples/ChangeDetection/RatioChDet.cxx
+ ${OTB_SOURCE_DIR}/Examples/ChangeDetection/CorrelChDet.cxx
+ ${OTB_SOURCE_DIR}/Examples/Classification/KdTreeBasedKMeansClustering.cxx
+ ${OTB_SOURCE_DIR}/Examples/Classification/ScalarImageKmeansModelEstimator.cxx
+ ${OTB_SOURCE_DIR}/Examples/Classification/ScalarImageKmeansClassifier.cxx
+ ${OTB_SOURCE_DIR}/Examples/Classification/BayesianPluginClassifier.cxx
+ ${OTB_SOURCE_DIR}/Examples/Classification/ExpectationMaximizationMixtureModelEstimator.cxx
+ ${OTB_SOURCE_DIR}/Examples/Classification/ScalarImageMarkovRandomField1.cxx
${OTB_SOURCE_DIR}/Examples/Learning/SVMPointSetModelEstimatorExample.cxx
${OTB_SOURCE_DIR}/Examples/Learning/SVMPointSetClassificationExample.cxx
${OTB_SOURCE_DIR}/Examples/Learning/SVMImageModelEstimatorExample.cxx
${OTB_SOURCE_DIR}/Examples/Learning/SVMImageClassificationExample.cxx
${OTB_SOURCE_DIR}/Examples/Learning/SVMImageEstimatorClassificationMultiExample.cxx
- ${OTB_SOURCE_DIR}/Examples/Visu/GreyVisuExample.cxx
+ ${OTB_SOURCE_DIR}/Examples/Visu/VisuExample1.cxx
)
diff --git a/SoftwareGuide/Latex/ChangeDetection.tex b/SoftwareGuide/Latex/ChangeDetection.tex
index 1f7e850b25..b3ca267c81 100644
--- a/SoftwareGuide/Latex/ChangeDetection.tex
+++ b/SoftwareGuide/Latex/ChangeDetection.tex
@@ -194,28 +194,29 @@ expression is actually used:
\section{Statistical Detectors}
\label{sec:StatisticalDetectors}
-%% \subsection{Local Correlation}
-
-%% The correlation coefficient measures the likelihood of a linear
-%% relationship between two random variables:
-%% \begin{equation}
-%% \begin{split}
-%% I_\rho(i,j) &= \frac{1}{N}\frac{\sum_{i,j}(I_1(i,j)-m_{I_1})(I_2(i,j)-m_{I_2})}{\sigma_{I_1}
-%% \sigma_{I_2}}\\
-%% & = \sum_{(I_1(i,j),I_2(i,j))}\frac{(I_1(i,j)-m_{I_1})(I_2(i,j)-m_{I_2})}{\sigma_{I_1}
-%% \sigma_{I_2}}p_{ij}
-%% \end{split}
-%% \end{equation}
+\subsection{Local Correlation}
+\label{sec:LocalCorrelation}
+The correlation coefficient measures the likelihood of a linear
+relationship between two random variables:
+\begin{equation}
+\begin{split}
+I_\rho(i,j) &= \frac{1}{N}\frac{\sum_{i,j}(I_1(i,j)-m_{I_1})(I_2(i,j)-m_{I_2})}{\sigma_{I_1}
+\sigma_{I_2}}\\
+& = \sum_{(I_1(i,j),I_2(i,j))}\frac{(I_1(i,j)-m_{I_1})(I_2(i,j)-m_{I_2})}{\sigma_{I_1}
+\sigma_{I_2}}p_{ij}
+\end{split}
+\end{equation}
-%% where $I_1(i,j)$ and $I_2(i,j)$ are the pixel values of the 2 images and
-%% $p_{ij}$ is the joint probability density. This is like using a linear model:
-%% \begin{equation}
-%% I_2(i,j) = (I_1(i,j)-m_{I_1})\frac{\sigma_{I_2}}{\sigma_{I_1}}+m_{I_2}
-%% \end{equation}
-%% for which we evaluate the likelihood with $p_{ij}$.
+where $I_1(i,j)$ and $I_2(i,j)$ are the pixel values of the 2 images and
+$p_{ij}$ is the joint probability density. This is like using a linear model:
+\begin{equation}
+I_2(i,j) = (I_1(i,j)-m_{I_1})\frac{\sigma_{I_2}}{\sigma_{I_1}}+m_{I_2}
+\end{equation}
+for which we evaluate the likelihood with $p_{ij}$.
-%% With respect to the difference detector, this one will be robust to
-%% illumination changes.
+With respect to the difference detector, this one will be robust to
+illumination changes.
+\input{CorrelChDet.tex}
%% \subsection{Mutual Information}
diff --git a/SoftwareGuide/Latex/Classification.tex b/SoftwareGuide/Latex/Classification.tex
index 4b9db4ba1c..959c599f20 100644
--- a/SoftwareGuide/Latex/Classification.tex
+++ b/SoftwareGuide/Latex/Classification.tex
@@ -1,10 +1,218 @@
\chapter{Classification}
\section{Introduction}
-What is feature extraction
+
+In statistical classification, each object is represented by $d$ features (a
+measurement vector), and the goal of classification becomes finding compact and
+disjoint regions (decision regions\cite{Duda2000}) for classes in a
+$d$-dimensional feature space. Such decision regions are defined by decision
+rules that are known or can be trained. The simplest configuration of a
+classification consists of a decision rule and multiple membership functions;
+each membership function represents a class. Figure~\ref{fig:simple}
+illustrates this general framework.
+
+\begin{figure}[h]
+ \centering
+ \includegraphics[width=0.7\textwidth]{DudaClassifier.eps}
+ \itkcaption[Simple conceptual classifier]{Simple conceptual classifier.}
+ \label{fig:simple}
+\end{figure}
+
+This framework closely follows that of Duda and
+Hart\cite{Duda2000}. The classification process can be described
+as follows:
+
+\begin{enumerate}
+\item{A measurement vector is input to each membership function.}
+\item{Membership functions feed the membership scores to the
+ decision rule.}
+\item{A decision rule compares the membership scores and returns a
+ class label.}
+\end{enumerate}
+
+\begin{figure}
+ \centering
+ \includegraphics[width=0.7\textwidth]{StatisticalClassificationFramework.eps}
+ \itkcaption[Statistical classification framework]{Statistical classification
+framework.}
+ \protect\label{fig:StatisticalClassificationFramework}
+\end{figure}
+
+This simple configuration can be used to formulated various classification
+tasks by using different membership functions and incorporating task specific
+requirements and prior knowledge into the decision rule. For example, instead
+of using probability density functions as membership functions, through
+distance functions and a minimum value decision rule (which assigns a class
+from the distance function that returns the smallest value) users can achieve a
+least squared error classifier. As another example, users can add a rejection
+scheme to the decision rule so that even in a situation where the membership
+scores suggest a ``winner'', a measurement vector can be flagged as ill
+defined. Such a rejection scheme can avoid risks of assigning a class label
+without a proper win margin.
+
+\subsection{k-d Tree Based k-Means Clustering}
+\label{sec:KdTreeBasedKMeansClustering}
+\ifitkFullVersion
+\input{KdTreeBasedKMeansClustering.tex}
+\fi
+
+\subsection{K-Means Classification}
+\label{sec:KMeansClassifier}
+\ifitkFullVersion
+\input{ScalarImageKmeansClassifier.tex}
+\fi
+\ifitkFullVersion
+\input{ScalarImageKmeansModelEstimator.tex}
+\fi
+
+\subsection{Bayesian Plug-In Classifier}
+\label{sec:BayesianPluginClassifier}
+
+\ifitkFullVersion
+\input{BayesianPluginClassifier.tex}
+\fi
+
+
+\subsection{Expectation Maximization Mixture Model Estimation}
+\label{sec:ExpectationMaximizationMixtureModelEstimation}
+
+\ifitkFullVersion
+\input{ExpectationMaximizationMixtureModelEstimator.tex}
+\fi
+
+\subsection{Classification using Markov Random Field}
+\label{sec:MarkovRandomField}
+
+Markov Random Fields are probabilistic models that use the correlation between
+pixels in a neighborhood to decide the object region. The
+\subdoxygen{itk::Statistics}{MRFImageFilter} uses the maximum a posteriori (MAP)
+estimates for modeling the MRF. The object traverses the data set and uses the
+model generated by the Mahalanobis distance classifier to get the the distance
+between each pixel in the data set to a set of known classes, updates the
+distances by evaluating the influence of its neighboring pixels (based on a MRF
+model) and finally, classifies each pixel to the class which has the minimum
+distance to that pixel (taking the neighborhood influence under consideration).
+The energy function minimization is done using the iterated conditional modes
+(ICM) algorithm \cite{Besag1986}.
+
+\ifitkFullVersion
+\input{ScalarImageMarkovRandomField1.tex}
+\fi
+
+
+
\section{Support Vector Machines}
\label{sec:SupportVectorMachines}
+
+Kernel based learning methods in general and the Support Vector
+Machines (SVM) in particular, have been introduced in the last years
+in learning theory for classification and regression tasks,
+\cite{vapnik}. SVM have been successfully applied to text
+categorization, \cite{joachims}, and face recognition,
+\cite{osuna}. Recently, they have been successfully used for the
+classification of hyperspectral remote-sensing images, \cite{bruzzoneSVM}.
+
+Simply stated, the approach consists in searching for the separating
+surface between 2 classes by the determination of the subset of
+training samples which best describes the boundary between the 2
+classes. These samples are called support vectors and completely
+define the classification system. In the case where the two classes are
+nonlinearly separable, the method uses a kernel expansion in order to make
+projections of the feature space onto higher dimensionality spaces
+where the separation of the classes becomes linear.
+
+ \subsection{Mathematical formulation}
+
+ This section reminds the basic principles of SVM learning and
+ classification. A good tutorial on SVM can be found in, \cite{burges}.
+
+We have $N$ samples represented by the couple $(y_i,\mathbf{x}_i),
+i=1\ldots N$ where $y_i \in \{-1,+1\}$ is the class label and
+$\mathbf{x}_i \in \mathbb{R}^n$ is the feature vector of dimension
+$n$. A classifier is a function $$f(\mathbf{x},\boldsymbol{\alpha}) :
+\mathbf{x}\mapsto y$$ where $\boldsymbol{\alpha}$ are the classifier
+parameters. The SVM finds the optimal separating hyperplane which
+fulfills the following constraints :
+ \begin{itemize}
+ \item The samples with labels $+1$ and $-1$ are on different
+ sides of the hyperplane.
+ \item The distance of the closest vectors to the hyperplane is
+ maximised. These are the support vectors (SV) and this distance is
+ called the margin.
+ \end{itemize}
+
+ The separating hyperplane has the equation
+ $$\mathbf{w}\cdot\mathbf{x}+b=0;$$ with $\mathbf{w}$ being its
+ normal vector and $x$ being any point of the hyperplane. The
+ orthogonal distance to the origin is given by
+ $\frac{|b|}{\|\mathbf{w}\|}$. Vectors located outside the
+ hyperplane have either $\mathbf{w}\cdot\mathbf{x}+b>0$ or
+ $\mathbf{w}\cdot\mathbf{x}+b<0$.
+
+ Therefore, the classifier function can be written as
+ $$f(\mathbf{x},\mathbf{w}, b)=sgn(\mathbf{w}\cdot\mathbf{x}+b).$$
+
+The SVs are placed on two hyperplanes which are parallel to the
+ optimal separating one. In order to find the optimal
+ hyperplane, one sets $\mathbf{w}$ and
+ $b$ : $$\mathbf{w}\cdot\mathbf{x}+b=\pm 1.$$
+
+Since there must not be any vector inside the margin, the following
+constraint can be used:
+ $$\mathbf{w}\cdot\mathbf{x}_i+b\ge +1\text{ if }y_i=+1;$$
+ $$\mathbf{w}\cdot\mathbf{x}_i+b\le -1\text{ if }y_i=-1;$$ which
+ can be rewritten as $$y_i(\mathbf{w}\cdot\mathbf{x}_i+b)-1\ge 0~ ~ \forall i.$$
+
+ The orthogonal distances of the 2 parallel hyperplanes to the
+ origin are $\frac{|1-b|}{\|\mathbf{w}\|}$ and
+ $\frac{|-1-b|}{\|\mathbf{w}\|}$. Therefore the modulus of the
+ margin is equal to $\frac{2}{\|\mathbf{w}\|}$ and it has to be
+ maximised.
+
+ Thus, the problem to be solved is:
+
+ \begin{itemize}
+ \item Find $\mathbf{w}$ and $b$ which minimise
+ $\left\{ \frac{1}{2}\|\mathbf{w}\|^2 \right\}$
+ \item under the constraint :
+ $y_i(\mathbf{w}\cdot\mathbf{x}_i+b)\ge 1~ ~ i=1\ldots N.$
+ \end{itemize}
+
+ This problem can be solved by using the Lagrange multipliers
+ with one multiplier per sample. It can be shown that only the
+ support vectors will have a positive Lagrange multiplier.
+
+ In the case where the two classes are not exactly linearly
+ separable, one can modify the constraints above by using
+ $$\mathbf{w}\cdot\mathbf{x}_i+b\ge +1 - \xi_i \text{ if }y_i=+1;$$
+ $$\mathbf{w}\cdot\mathbf{x}_i+b\le -1+\xi_i \text{ if }y_i=-1;$$
+ $$\xi_i\ge 0~ ~\forall i.$$
+
+ If $\xi_i > 1$, one considers that the sample is wrong. The
+ function which has then to be minimised is
+ $\frac{1}{2}\|\mathbf{w}\|^2 + C\left( \sum_i \xi_i\right); $,
+ where $C$ is a tolerance parameter. The optimisation problem
+ is the same than in the linear case, but one multiplier has to
+ be added for each new constraint $\xi_i\ge 0$.
+
+ If the decision surface needs to be non-linear, this solution
+ cannot be applied and the kernel approach has to be adopted.
+
+
+One drawback of the SVM is that, in their basic version, they can only
+solve two-class problems. Some works exist in the field of multi-class
+SVM (see \cite{allwein00reducing,weston98multiclass}, and the
+comparison made by \cite{hsu01comparison}), but they are
+not used in our system.
+
+For problems with $N > 2$ classes, one can choose either to train $N$
+SVM (one class against all the others), or to train $N\times(N-1)$ SVM
+(one class against each of the others). In the second approach, which
+is the one that we use, the final decision is taken by choosing the
+class which is most often selected by the whole set of SVM.
+
+
\subsection{Learning With PointSets}
\label{sec:LearningWithPointSets}
\input{SVMPointSetModelEstimatorExample}
diff --git a/SoftwareGuide/Latex/SoftwareGuide.tex b/SoftwareGuide/Latex/SoftwareGuide.tex
index 2ccecc66a0..f7bcbb7b7e 100644
--- a/SoftwareGuide/Latex/SoftwareGuide.tex
+++ b/SoftwareGuide/Latex/SoftwareGuide.tex
@@ -64,6 +64,7 @@ colorlinks,linkcolor={blue},citecolor={blue},urlcolor={blue},
]{hyperref}
\fi
+\usepackage{amsmath,amssymb,amsfonts}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
%
diff --git a/SoftwareGuide/Latex/Visualization.tex b/SoftwareGuide/Latex/Visualization.tex
index 3281ba90c8..e2147cd14c 100644
--- a/SoftwareGuide/Latex/Visualization.tex
+++ b/SoftwareGuide/Latex/Visualization.tex
@@ -1,20 +1,21 @@
\chapter{Image Visualization}
\label{chap:ImageVisualization}
-Even if OTB is not a visualization toolkit as for instance
-\url{VTK}{http://www.vtk.ork}, some simple functionnalities for image
-visualization are given in the toolbox. Indeed, for algorithm
-prototyping, it is sometimes more useful to \emph{see} the result on
-the screen, than saving it to a file and then open in with an external
-viewer.
+Even if OTB is not a visualization toolkit as for instance VTK
+(\emph{The Visualization Toolkit} \url{http://www.vtk.org}), some
+simple functionnalities for image visualization are given in the
+toolbox. Indeed, for algorithm prototyping, it is sometimes more
+useful to \emph{see} the result on the screen, than saving it to a
+file and then open in with an external viewer.
OTB provides the \doxygen{otb::ImageViewer} class which is compatible
with the pipeline and can therefore replace the
\doxygen{otb::ImageFileWriter} during proto-typing phases.
-\section{Viewing Grey Level Images}
-\label{sec:ViewingGreyLevelImages}
-\input{GreyVisuExample.tex}
-\section{Viewing RGB Images}
-\label{sec:ViewingRGBImages}
-\section{Viewing Multiband Images}
-\label{sec:ViewingMultibandImages}
+\input{VisuExample1.tex}
+%% \section{Viewing Grey Level Images}
+%% \label{sec:ViewingGreyLevelImages}
+
+%% \section{Viewing RGB Images}
+%% \label{sec:ViewingRGBImages}
+%% \section{Viewing Multiband Images}
+%% \label{sec:ViewingMultibandImages}
--
GitLab