itksample.h

InsightToolkit-1.4.0(有大量的优化算法程序)

/*=========================================================================

  Program:   Insight Segmentation & Registration Toolkit
  Module:    $RCSfile: itkSample.h,v $
  Language:  C++
  Date:      $Date: 2003/09/10 14:29:47 $
  Version:   $Revision: 1.17 $

  Copyright (c) Insight Software Consortium. All rights reserved.
  See ITKCopyright.txt or http://www.itk.org/HTML/Copyright.htm for details.

     This software is distributed WITHOUT ANY WARRANTY; without even 
     the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR 
     PURPOSE.  See the above copyright notices for more information.

=========================================================================*/
#ifndef __itkSample_h
#define __itkSample_h

#include <vector>

#include "itkMacro.h"
#include "itkPoint.h"
#include "itkSize.h"
#include "itkObject.h"
#include "itkFixedArray.h"

namespace itk{ 
namespace Statistics{

/** \class Sample 
 *  \brief Sample defines common iterfaces for each subclasses
 *
 * A measurement vector is a vector of different types of values. For example,
 * users might have an image with each pixel have a gray level intensity value
 * and a gradient magnitude. In this example, user think each pixel as a 
 * measurement vector that has two types of values, intensity and gradient
 * magnitude. In this sense, the meaning of "measurement" can be interchaneable 
 * with "measurement".
 * 
 * The primary purpose of Sample clases are storing such 
 * measurement vectors and providing different types of element access methods.
 *
 * The "TMeasurement" template argument defines the type of each values
 * in a measurement vector. The "VMeasurementVectorSize" template argument 
 * defines 
 * the number of elements in a measurement vector. The type of measurement 
 * vector 
 * is fixed as ITK Point. The resulting measurement vector type is 
 * Point< TMeasurement, VMeasurementVectorSize >.
 *
 * There are two common ways of accessing data. First, users can use 
 * an "InstanceIdentifier", say 'id' which is sequential index of 
 * measurement vectors to access a measurement vector in a container. 
 * Second, users can access a measurement value using dimension and 
 * the index of a measurement value, say 'n' in that dimension. The 'id' and 'n'
 * is equal in the ListSample classes. However, in Histogram class, its
 * different. See Histogram. For algorithms using these indexes, there are
 * methods to get the length of each dimension, GetSize() and 
 * GetSize(dimension). See each method's comment.
 *
 * Users may need to know two other groups of methods to fully utilize
 * Sample. One group of methods is about frequency of measurement
 * vectors. In Histogram classes, you can use such methods to get/set the
 * frequency of each measurement vector. The other group of methods is about
 * characteristics of each subclasses of this class. If user want to 
 * create or know the characteristics of a subclass, use them. The set of
 * characteristics tell that if the data is sorted (true for Histogram),
 * if the each measurement vector is unique (true for Histogram), and if the
 * container supports "meaningful" frequency interfaces. In ListSample
 * classes, users cannot set frequency, and GetFrequency() methods return 
 * one if the element exists, else zero. So, currently ListSample's
 * supporting frequency is set to false. 
 */

template < class TMeasurementVector >
class ITK_EXPORT Sample : public Object
{
public:
  /** Standard class typedefs */
  typedef Sample  Self;  
  typedef Object Superclass ;
  typedef SmartPointer< Self > Pointer ;

  /** Run-time type information (and related methods) */
  itkTypeMacro(Sample, Object);

  /** MeasurementVector typedef support */ 
  typedef TMeasurementVector MeasurementVectorType ;

  /** typedef for each element in a measurement vector*/ 
  typedef typename TMeasurementVector::ValueType MeasurementType ;

  /** frequency value typedef*/
  typedef float FrequencyType ;

  /** Instance Identifier typedef support
   * this identifier will be unique sequential id for each measurement vector
   * in a Sample subclass.*/ 
  typedef unsigned long InstanceIdentifier ;

  itkStaticConstMacro( MeasurementVectorSize, unsigned int,
                       TMeasurementVector::Length ) ;

  virtual unsigned int Size() const = 0 ;

  /** returns the measurement of the instance which is identified by the 'id'*/
  virtual MeasurementVectorType GetMeasurementVector(const InstanceIdentifier &id) = 0 ;

  /** returns the frequency of the instance which is identified by the 'id'*/
  virtual FrequencyType GetFrequency(const InstanceIdentifier &id) const = 0 ;

  /** returns the total frequency for the 'd' dimension*/
  virtual FrequencyType GetTotalFrequency() const 
    = 0 ;

  
protected:
  Sample() {}
  virtual ~Sample() {}
  void PrintSelf(std::ostream& os, Indent indent) const
  {
    Superclass::PrintSelf(os,indent);
  }

  
private:
  Sample(const Self&) ; //purposely not implemented
  void operator=(const Self&) ; //purposely not implemented
} ; // end of class

} // end of namespace Statistics 
} // end of namespace itk

#endif