Index changes of seed points in ITK original image after ROI and downsampling

Original image in vtk Physical coordinates and index Sort out the relationship between ：

1、vtk Interaction ： stay vtk From the slice world point, use Image->TransformPhysicalPointToIndex(worldpoint) Function can be used to get the index;

2、ITk At the same position of the image index Extracting ROI Area and Downsampling What will become after ？ How to calculate ？

The following functions are extracted by me ROI Area , Then on ROI Code for area downsampling ：

#include "itkRegionOfInterestImageFilter.h"
#include "itkResampleImageFilter.h"

template <typename ImageType>
void getROIandReshampe(ImageType* inputImage,ImageType* outputImage)
{
    auto size = inputImage->GetLargestPossibleRegion().GetSize();

    typedef itk::RegionOfInterestImageFilter<ImageType,ImageType> FilterType;
    FilterType::Pointer roi = FilterType::New();

   ImageType::IndexType start;
    start[0] = 50;
    start[1] = 50;
    start[2] = 10;
   ImageType::SizeType roiSize;
    roiSize[0] = 200;
    roiSize[1] = 240;
    roiSize[2] = 90;

   ImageType::RegionType desiredRegion;
    desiredRegion.SetSize(roiSize);
    desiredRegion.SetIndex(start);

    roi->SetRegionOfInterest(desiredRegion);
    roi->SetInput(inputImage);
    roi->Update();
 
    auto spacing = inputImage->GetSpacing();

    int a = 5;
    spacing[0] = a;
    spacing[1] = a;
    spacing[2] = a;
    typedef itk::ResampleImageFilter<ImageType,ImageType> ResampleImageFilterType;
    ResampleImageFilterType::Pointer resampleFilter = ResampleImageFilterType::New();
    resampleFilter->SetInput(inputImage);

    typedef itk::NearestNeighborInterpolateImageFunction<ImageType, double> InterpolatorType;
    InterpolatorType::Pointer interpolator = InterpolatorType::New();
    resampleFilter->SetInterpolator(interpolator);
    resampleFilter->SetDefaultPixelValue(0);
    resampleFilter->SetOutputSpacing(spacing);
    resampleFilter->SetSize(inputImage->GetLargestPossibleRegion().GetSize());
    resampleFilter->SetOutputOrigin(inputImage->GetOrigin());
    resampleFilter->SetOutputDirection(inputImage->GetDirection());
    resampleFilter->Update();
  
    outputImage = resampleFilter->GetOutput();

}

a key ：

itk Of ROI Function to get the image size Same as the original drawing size The same big as that , therefore ROI In the same position index No change ;

for example ： A seed dot in the original picture index by ：p : [px,py,pz]

ROI This point is included in ROI In the region ： hypothesis ROI The starting point of index by ROIstart, The size of the area is ROISize

                                                  Seeded index Still ：[px,py,pz]

After downsampling ： The size of the whole physical space of the obtained image has not changed , change spacing after , The physical location at the same pixel location （ World coordinates ） Same as the original , however index atypism , Because of the new spacing After getting bigger , The position between pixels becomes larger ,
p It's from the original index;
spacing、origin Is the parameter of the original drawing ;
space Is the parameter of downsampling ;

So the seed order index Turn into ：
shampleIndex[0] = ((px - start[0]) * spacing[0] + origin[0]) / space[0];

shampleIndex[1] = ((py - start[1]) * spacing[1] + origin[1]) / space[1];

shampleIndex[2] = ((pz - start[2]) * spacing[2] + origin[2]) / space[2];
At this point, the seed points of the original graph are obtained index In the downsampled image Of index

 

 

Revision summary