ITK入门教程（3）仅对图像的指定区域应用过滤器

1.函数说明

template<typename TPixel, unsigned int VImageDimension = 2> class Image : public itk::ImageBase<VImageDimension>模板化的 n 维图像类。

图像基于像素类型（对因变量建模）和维度（自变量的数量）进行模板化。像素数据的容器是 ImportImageContainer。

在像素容器内，图像被建模为数组，由起始索引和大小定义。

Image 的超类 ImageBase 根据图像在物理空间中的位置、图像在物理空间中的方向、像素的大小以及图像本身的范围来定义图像的几何形状。 ImageBase 提供了在索引和物理空间坐标系之间转换的方法。

像素可以使用 SetPixel() 和 GetPixel() 方法直接访问，也可以通过定义它们遍历的图像区域的迭代器访问。

使用 SetRequestedRegion 告诉过滤器只对指定的 ImageRegion 进行操作的关键是在过滤器 GetOutput() 上调用 SetRequestedRegion。也就是说，要告诉 DerivativeImageFilter 只对一个小区域进行操作，你必须这样做

derivativeFilter->GetOutput()->SetRequestedRegion(smallRegion);
derivativeFilter->Update();

2.代码展示

（1）Python代码

#!/usr/bin/env python

import itk

Dimension = 2
PixelType = itk.F
ImageType = itk.Image[PixelType, Dimension]

smallSize = itk.Size[Dimension]()
smallSize.Fill(10)

index = itk.Index[Dimension]()
index.Fill(0)

region = itk.ImageRegion[Dimension]()
region.SetIndex(index)
region.SetSize(smallSize)

bigSize = itk.Size[Dimension]()
bigSize.Fill(10)

RandomSourceType = itk.RandomImageSource[ImageType]
randomImageSource = RandomSourceType.New()
randomImageSource.SetNumberOfWorkUnits(1)  # to produce non-random results
randomImageSource.SetSize(bigSize)
randomImageSource.GetOutput().SetRequestedRegion(region)

print("Created random image.")

DerivativeImageFilterType = itk.DerivativeImageFilter[ImageType, ImageType]

derivativeFilter = DerivativeImageFilterType.New()
derivativeFilter.SetInput(randomImageSource.GetOutput())
derivativeFilter.SetDirection(0)  # "x" axis
derivativeFilter.GetOutput().SetRequestedRegion(region)
derivativeFilter.Update()

print("Computed derivative.")

（2）C++代码

#include "itkImage.h"
#include "itkRandomImageSource.h"
#include "itkDerivativeImageFilter.h"

int main(int, char *[])
{
  constexpr unsigned int Dimension = 2;
  using PixelType = float;

  using ImageType = itk::Image<PixelType, Dimension>;

  ImageType::SizeType smallSize;
  smallSize.Fill(10);

  ImageType::IndexType index;
  index.Fill(0);

  ImageType::RegionType region(index, smallSize);

  ImageType::SizeType bigSize;
  bigSize.Fill(10000);

  using RandomSourceType = itk::RandomImageSource<ImageType>;
  RandomSourceType::Pointer randomImageSource = RandomSourceType::New();
  randomImageSource->SetNumberOfWorkUnits(1); // to produce non-random results
  randomImageSource->SetSize(bigSize);
  randomImageSource->GetOutput()->SetRequestedRegion(smallSize);

  std::cout << "Created random image." << std::endl;

  using DerivativeImageFilterType = itk::DerivativeImageFilter<ImageType, ImageType>;

  DerivativeImageFilterType::Pointer derivativeFilter = DerivativeImageFilterType::New();
  derivativeFilter->SetInput(randomImageSource->GetOutput());
  derivativeFilter->SetDirection(0); // "x" axis
  derivativeFilter->GetOutput()->SetRequestedRegion(smallSize);
  derivativeFilter->Update();

  std::cout << "Computed derivative." << std::endl;

  return EXIT_SUCCESS;
}

参考目录
https://itk.org/ITKExamples/src/Core/Common/ApplyAFilterOnlyToASpecifiedRegionOfAnImage/Documentation.html