/** 
 \defgroup DataRepresentation Data Representation Objects
 The Insight Toolkit includes several data representation objects such as 
 Image, Mesh, Pixel, Points, etc. 
*/

/**
  \defgroup ImageObjects Image Representation Objects
  \ingroup DataRepresentation
  This category includes the objects required to represent images in ITK. 
*/

/**
  \defgroup MeshObjects Mesh Representation Objects
  \ingroup DataRepresentation
  This category includes the objects required to represent meshes in ITK.
*/

/**
  \defgroup PathObjects  Path Representation Objects
  \ingroup DataRepresentation
  This category includes the objects required to represent paths in ITK.
*/

/**
  \defgroup Geometry Geometry Representation Objects
  \ingroup DataRepresentation

  This category include the objects required to represent geometrical entities
  like positions, vectors and space mappings.
   
  A detailed description of the rationale for these classes can be found in
  \ref GeometryPage  
*/
 

/** 
 \defgroup DataAccess Data Access Objects

 The Insight Toolkit includes several ways to access data through the user of 
 iterators, pointers, indexes, etc.
*/

/**
  \defgroup TensorObjects  Objects Related to Tensor Images

  This category includes the objects required for representing diffusion tensor images in ITK.
*/


/**
\defgroup ImageAccess Image Access Objects
\ingroup DataAccess
*/


/**
\defgroup MeshAccess Mesh Access Objects
\ingroup DataAccess
*/


/**
\defgroup Iterators Iterators
\ingroup DataAccess
Iterators are the mechanism used to walk over the content of a particular data object.
They allow to define paths and directions along which the data should be walked through.
*/

/**
\defgroup ImageIterators Image Iterators
\ingroup Iterators
Image Iterators allow to go through the content of an image in a predefined way.
For a detailed description of iterators rationale see \ref ImageIteratorsPage
*/




/** 
 \defgroup DataProcessing Data Processing Objects

 The Insight Toolkit includes several ways to process the data using objects such as adaptors, functions, filters, and
 transforms.  
*/


/**
  \defgroup Filters Filters
  \ingroup DataProcessing
  Filters implement the operations on the pipeline architecture.
*/

/**
  \defgroup ImageFilters Image Filters
  \ingroup Filters

  Image filters process input images and produce output images. Inputs are
  unmodified. The pipeline architecture makes provisions for supporting
  streaming by using packets of data defined by regions
  \sa Image
  \sa PhysicalImage
  \sa ImageRegion
*/

/**
  \defgroup MeshFilters Mesh Filters
  \ingroup Filters

  Mesh filters process input meshes and produce output meshes. Inputs are
  unmodified. 
  \sa Mesh
*/

/**
  \defgroup IntensityImageFilters Intensity Image Filters 
  \ingroup ImageFilters

  Intensity Image filters only alter the values stored in image pixels.
  \sa Image
  \sa PhysicalImage
  \sa ImageRegion
*/

/**
  \defgroup MathematicalMorphologyImageFilters Mathematical Morphology Image Filters 
  \ingroup IntensityImageFilters

  Mathematical morphology filters are a particular class of cellular automata.
  They modify the value of a pixel based on the values of a neighborhood. 
  The neighborhood is now as the structured element.

  \sa Image
  \sa PhysicalImage
  \sa ImageRegion
  \sa BinaryMorphologicalFilterBase
*/


/**
 \defgroup ImageEnhancement Image Enhancement Filters
 \ingroup ImageFilters
 Image enhancement filters process an image to enhance the appearance 
 of an image either for visualization purposes or for further processing.
 Examples of image enhancement filters available in ITK are: anisotropic diffusion,
 Gaussian filter, and histogram equalization.
*/  

/**
 \defgroup ImageFeatureExtraction Image Feature Extraction Filters
 \ingroup ImageFilters
 Image feature extraction filters process an image to extract features of interest
 such as gradients, edges, distances, etc. 
 Examples of image feature extraction algorithms available in ITK are: image gradients, 
 first and second derivatives, and Danielson distance.
*/  

/**
 \defgroup GradientFilters Image Gradient Filters
 \ingroup ImageFeatureExtraction
 These filters compute local gradients of images. 
*/  


/**
 \defgroup GeometricTransforms Geometric Transformation Filters
 \ingroup Filters
 Geometric transformation Filters transform the coordinates of an image in various ways. 
 Examples of geometric transformation Filters available in ITK are: image shrinking, and 
 affine transformation.
*/ 

/**
 \defgroup PyramidImageFilter Image Pyramid Filters
 \ingroup GeometricTransforms
 Image pyramid filters generate a set of downsampled versions of an image according
 to a user defined multi-resolution schedule.
*/ 

/**
 \defgroup ImageSegmentation Image Segmentation Filters
 \ingroup ImageFilters
 Image segmentation filters process an image to 
 partition it into meaningful regions. Various types of image segmentation algorithms 
 are available in ITK - some examples are, unsupervised pixel classification methods, region-growing methods, 
 watershed-based methods, deformable-model based methods, and level-set based methods. 
*/  

/**
 \defgroup IntensityImageSegmentation Intensity-Based Image Segmentation Filters
 \ingroup ImageSegmentation
 Intensity based image segmentation filters use intensity values  
 of image pixels to segment an image. Typically, spatial contiguity is not considered in 
 intensity-based segmentation filters. 
 Examples of intensity-based algorithms in ITK are supervised and unsupervised pixel classification algorithms.
*/  

/**
 \defgroup ClassificationFilters Pixel Classification Filters
 \ingroup IntensityImageSegmentation
  Pixel classification filters use statistical classification algorithms to assign a label to 
a given image pixel. Classification algorithms can be supervised when training data is available 
or unsupervised when no training data is available. 
*/  

/**
 \defgroup SupervisedClassificationFilters Supervised Classification Filters
 \ingroup ClassificationFilters
 Supervised classification filters rely on the existence of training data to classify 
 pixels into different types. An example of supervised classification algorithm in ITK is
 the Gaussian classifier that uses the training data to build Gaussian models of intensity distributions.
*/  

/**
 \defgroup UnSupervisedClassificationFilters Unsupervised Classification Filters
 \ingroup ClassificationFilters
 Unsupervised classification filters typically cluster the image intensity data into different groups. 
 An example of unsupervised classification algorithm in ITK is the K-Means clustering algorithm. 
*/


/**
 \defgroup WatershedSegmentation Watershed-based Segmentation Filters
 \ingroup IntensityImageSegmentation
 These filters segment an image based on intensity values using the watershed algorithm.  
*/  

/**
 \defgroup RegionBasedSegmentation Region-Based Segmentation Filters
 \ingroup ImageSegmentation
 These filters segment an image based on similarity of intensity values between spatially adjacent 
 pixels. Examples of region-based segmentation filters in ITK include fuzzy connectedness, region growing, and 
 Markov Random Fields.
*/

/**
 \defgroup FuzzyConnectednessSegmentation  Fuzzy Connectedness-based Segmentation Filters
 \ingroup RegionBasedSegmentation
 These filters segment an image based on fuzzy connectedness principles. Here you typically 
start with one or more seed points and grow regions around these seed points based on fuzzy affinity.
*/  

/**
 \defgroup RegionGrowingSegmentation Region Growing Filters
 \ingroup RegionBasedSegmentation
Typically region growing involves starting several small regions on an image and merging them iteratively based on some 
pixel intensity similarity criterion. ITK provides an intensity and edge-based region-growing algorithm for segmentation.
*/

/**
 \defgroup MRFFilters Markov Random Field-based Filters
 \ingroup RegionBasedSegmentation
Markov Random Field (MRF)-based Filters assume that the segmented image is Markovian in nature, i.e., adjacent 
pixels are likely to be of the same class. These methods typically combine intensity-based Filters with MRF prior 
models also known as Gibbs prior models.
*/


/**
 \defgroup ModelImageSegmentation Model-Based Image Segmentation Filters
 \ingroup ImageSegmentation
 These filters segment an image by starting with a model and then updating the model based on 
image features and the updates are typically constrained by a-priori knowledge about the models. 
Examples of these types of algorithms in ITK include: deformable model (snakes)-based algorithms and level set-based algorithms.
*/  

/**
 \defgroup MeshSegmentation Mesh Segmentation Filters
 \ingroup ModelImageSegmentation
These algorithms represent models using a mesh and update the models based on image features. 
Examples of this type of filter in ITK include: balloon force filter and the deformable mesh filter.
*/  

/**
 \defgroup LevelSetSegmentation Level Set-Based Segmentation Filters
 \ingroup ModelImageSegmentation
 These algorithms represent models implicitly using level-sets and update the models based on image features. 
Examples of these types of segmentation methods in ITK include: curvature flow-based filters, 
  fast marching filters, and shape-detection filters.
*/  

/**
 \defgroup HybridSegmentation Hybrid Segmentation Filters
 \ingroup ImageSegmentation
 These filters are hybrid between intensity-based, region-based, or model-based image
 segmentation filters. 
*/  


/**
 \defgroup RegistrationFilters Registration Filters
 \ingroup Filters
*/  

/**
 \defgroup RegistrationComponents Components of Registration Methods
 Registration methods are implemented by combining basic components. This framework allows
 great flexibility in the construction of registration methods, and maximize code reuse.
 The basic components of a registration method are described in \ref RegistrationPage
 \ingroup RegistrationFilters
*/  

/**
 \defgroup RegistrationMetrics Similarity Metrics of Registration Methods
 Similarity metrics are the objects that evaluate how similar two objects are. They are
 used in registration to quantify how well a transform is mapping the reference object 
 on top of the target object.
 \ingroup RegistrationComponents

*/  


/**
 \defgroup ImageRegistration Image Registration Methods
 \ingroup RegistrationFilters
*/  

/**
 \defgroup RigidImageRegistration Rigid Registration Methods
 \ingroup ImageRegistration
*/  

/**
 \defgroup AffineImageRegistration Affine Registration Methods
 \ingroup ImageRegistration
*/  


/**
 \defgroup DeformableImageRegistration Deformable Registration Methods
 \ingroup ImageRegistration
*/  

/**
 \defgroup ModelImageRegistration Model - Image Registration Methods
 \ingroup RegistrationFilters
*/  

/**
 \defgroup PointSetToImageRegistration PointSet to Image Registration Methods
 \ingroup ModelImageRegistration
*/  


/**
\defgroup IOFilters Input and Output Filters
\ingroup Filters
*/



/**
\defgroup DataSources Data Sources
\ingroup DataProcessing
*/



/**
\defgroup Transforms Transforms 
\ingroup DataProcessing
*/


/**
\defgroup ImageAdaptors Image Adaptors 
\ingroup DataProcessing

Image Adaptors are an implementation of the <EM>Adaptor Design Pattern</EM>. They are 
designed to present an image of a particular type as being an image of a different type.
Adaptors perform simple operations on pixel values for which is not easy to justify the
use of an image filter.

One of the principal tasks of ImageAdaptors is to perform casting. 

For example: you have an image whose pixels are of type <TT>unsigned char</TT> and you 
need to feed this image in a process that expects pixels of type <TT>double</TT>. 
You have the option of using and ImageFilter that convert the input <TT>unsigned char </TT> image into another of pixel type <TT>double</TT>. 
However this filter will allocate memory for this second image and will need to be executed. 
Image Adaptors allow to simulate that you have made the conversion but will avoid the overhead in memory. 
There is however a penalty in performance. 

The mechanism used by image adaptors is to provide a simple function that will be used by ImageIterator (see \ref ImageIteratorsPage) to convert the value of a pixel, in a pixel-by-pixel basis.
*/


/**
\defgroup Functions Functions
\ingroup DataProcessing
Functions provide an efficient mechanism for computing values
*/

/**
\defgroup ImageFunctions Image Functions
\ingroup Functions
Image functions compute single values using data from a previously
specified Image. A value can be computed at an image index,
continuous index or point.

\sa Image
\sa Index
\sa ImageFunction
*/

/**
\defgroup ImageInterpolators Image Interpolators
\ingroup ImageFunctions
Image interpolators compute pixel values in non-grid positions. A value can be computed at an image index, continuous index or point.

\sa Image
\sa ImageFunction
*/


/**
\defgroup SpatialFunctions Spatial Functions
\ingroup Functions
*/

/**
\defgroup FiniteDifferenceFunctions Finite Difference Functions
\ingroup Functions
Finite difference functions are used in the finite difference solver (FDS)
framework for solving partial differential equations on images using
an iterative, finite difference update scheme.

\sa FiniteDifferenceImageFilter
\sa FiniteDifferenceFunction 
*/

/**
\defgroup Operators Operators
\ingroup DataProcessing
Operators implements the abstraction of performing an operation using data
from a neighborhood of a pixel. ITK Operators work in conjunction with
Neighborhood iterators in order to walk over an image.
*/



/**
  \defgroup Numerics Numerics
 
  Insight provides support for numerical operations at two levels. First, 
  Insight uses an external library called VNL, which is one component of
  the VXL toolkit. This library provides linear algebra, optimization, 
  and FFTs. Second, Insight provides numerical optimizers
  designed for the registration framework and statistical
  classes designed to be used for classification and segmentation.
 */


/**
  \defgroup Optimizers Optimizers
  \ingroup Numerics RegistrationComponents
  \sa RegistrationComponents
 
  Set of Optimization methods. Some of these methods are adaptors for 
  classes already defined in the VNL library. These methods have been
  particularly tailored to be used as modular components in the 
  registration framework.
 */

/**
\defgroup SystemObjects System Objects
*/

/**
\defgroup ITKSystemObjects ITK System Objects
\ingroup SystemObjects
These objects are the basic system objects used in building ITK.
*/

/**
\defgroup OSSystemObjects OS System Objects
\ingroup SystemObjects
 These objects are the basic operating system related system objects in ITK.
*/



/**
  \defgroup ThreadSafetyGroup  Thread Safety
  This groups catalog classes according to its compliance with thread safety.
  ITK is designed to be thread-safe, but in some particular cases this
  capability cannot be supported and for this reason a classification is
  needed.
*/  

/**
 \defgroup ThreadSafe Thread Safe classes
 \ingroup ThreadSafetyGroup
*/  

/**
 \defgroup ThreadUnSafe Thread Unsafe classes
 \ingroup ThreadSafetyGroup
*/  

/**
 \defgroup ThreadSafetyUnknown Thread Safety Unknown
 \ingroup ThreadSafetyGroup
*/



/**
  \defgroup MultithreadingGroup   Support for Multithreading
  This category classifies filters according to their support 
  for performing processing in multiple threadS
*/


/**
  \defgroup Multithreaded   Multithreaded Filters
  \ingroup MultithreadingGroup
  Filters on this category divide processing across several threads.
*/


/**
  \defgroup Singlethreaded  Singlethreaded Filters 
  \ingroup MultithreadingGroup
  Filter on this category perform all its processing in a single thread
*/


/**
  \defgroup ShouldBeThreaded Filters that can potentialy be modified to be Threaded 
  \ingroup MultithreadingGroup
  Filter on this category could be implemented to use several processing threads
*/


/** 
  \defgroup StreamingGroup     Processing images region by region 
  This category classifies filters according to their capatity for supporting
  Streaming. 
*/

/** 
  \defgroup Streamed      Filters supporting Streaming    
  \ingroup StreamingGroup
  Filter can respond to a request for a portion of the output using 
  only a portion of the input.
*/


/**
  \defgroup CannotBeStreamed  Filters that cannot be streamed      
  \ingroup StreamingGroup
  Filter requires either all the input or produces all the output.
*/


/**
  \defgroup ShouldBeStreamed  Filters that could be implemented to be streamed 
  \ingroup StreamingGroup
  Filters that can potentially be modified to support streaming but currently 
  are not supporting it.
*/




/**
  \defgroup Deprecated Deprecated classes that are scheduled to be removed from the Toolkit
  The classes on this group will be removed soon. Their funcionality is now provided 
  by other classes or changes in the toolkit have made them useless. Please report to
  their documentation and look in to their "Deprecated" section. This section should
  indicate what to do to replace this class in your code.
*/