Homec4science
Diffusion ShoulderDB_ImportSegmentMeasureAnalyze (feat_muscle_subdivisions) Browse

R8218/ShoulderCase/@MusclesContainer/createRotatorCuffSlicesMatthieu.m
fb70e1933436feat_muscle_subdivisions

Last Change
Raw File

createRotatorCuffSlicesMatthieu.m
View Options

function createRotatorCuffSlicesMatthieu(obj,sliceName)
% This function and its results might be reffered as
% Matthieu algorithm and Matthieu slices in some studies.
%
% This method has been written to simplify, improve, and make reusable
% the Nathan's slicing algorithm written in createRotatorCuffSlicesNathan().
% This means:
% - taking advantage of the MATLAB obliqueslice() method wich wasn't available
% when Nathan's algorithm has been written.
% - encapsulating the slicing data and methods in a DicomVolumeSlicer object.
% - being able to orient a slice according to dicom points coordinates.
% - being able to crop the slice according to dicom points coordinates
% and cropping dimensions given in the dicom set units.
%
% Images and pixel spacings are saved at:
% shoulder/dataDev/*/*/*/SCase-IPP/CT-SCase-IPP/matlab/shoulder/muscles/oblique_slice/
assert(not(obj.shoulder.scapula.coordSys.isEmpty),'Scapula coordinate system not measured.');
[imageForSegmentation,imageForMeasurements,imagesPixelSpacings,imagesPixelCoordinates] = getSlices(obj);
saveImages(obj,imageForSegmentation,imageForMeasurements,sliceName);
saveImagesPixelSpacings(obj,imagesPixelSpacings,sliceName);
saveImagesPixelCoordinates(obj,imagesPixelCoordinates,sliceName);
end
function [slice8Bit,slice16Bit,slicePixelSpacings,slicePixelCoordinates] = getSlices(obj)
SCase = obj.shoulder.SCase;
scapula = obj.shoulder.scapula;
path = SCase.dataDicomPath;
try
rotatorCuffSlicer = DicomVolumeSlicer(SCase.getSmoothDicomPath);
catch
rotatorCuffSlicer = DicomVolumeSlicer(SCase.dataDicomPath);
end
% The 3D volume must be normalised for the slices not to be distorted
rotatorCuffSlicer.setMinimalResolutionDividedBy(2);
rotatorCuffSlicer.normaliseVolume;
rotatorCuffSlicer.slice(scapula.coordSys.origin,scapula.coordSys.ML);
% The following lines set the slice to display a 300x300 mm area around the
% scapula with the "Y" shape of the scapula oriented up
rotatorCuffSlicer.orientSliceUpwardVector(scapula.coordSys.origin,scapula.coordSys.IS);
rotatorCuffSlicer.crop(scapula.coordSys.origin-100*scapula.coordSys.IS,300,300);
% Rotator cuff segmentation code requires 1024x1024 pixels images
if size(rotatorCuffSlicer.sliced,1) > size(rotatorCuffSlicer.sliced,2)
rotatorCuffSlicer.resize([1024 nan]);
else
rotatorCuffSlicer.resize([nan 1024]);
end
rotatorCuffSlicer.addEmptyBackgroundToSlice(1024,1024);
rawSlice = rotatorCuffSlicer.sliced;
rotatorCuffSlicer.rescaleSliceToUint8;
slice8Bit = rotatorCuffSlicer.sliced;
rotatorCuffSlicer.sliced = rawSlice;
rotatorCuffSlicer.rescaleSliceToInt16;
slice16Bit = rotatorCuffSlicer.sliced;
slicePixelCoordinates = rotatorCuffSlicer.getSlicedPixelCoordinates();
% Flip the slices for right and left shoulders to give similar slices
if obj.shoulder.side=='L'
slice8Bit=rot90(flipud(slice8Bit),2);
slice16Bit=rot90(flipud(slice16Bit),2);
slicePixelCoordinates=rot90(flipud(slicePixelCoordinates),2);
end
slicePixelSpacings = rotatorCuffSlicer.slicedPixelSpacings;
end
function saveImages(obj,imageForSegmentation,imageForMeasurements,sliceName)
imwrite(imageForSegmentation,...
fullfile(obj.slicesDataPath,sliceName + "_ForSegmentation.png"));
save(fullfile(obj.slicesDataPath,sliceName + "_ForMeasurements.mat"),...
'imageForMeasurements');
end
function saveImagesPixelSpacings(obj,imagesPixelSpacings,sliceName)
save(fullfile(obj.slicesDataPath,sliceName + "_PixelSpacings.mat"),...
'imagesPixelSpacings');
end
function saveImagesPixelCoordinates(obj,imagesPixelCoordinates,sliceName)
save(fullfile(obj.slicesDataPath,sliceName + "_PixelCoordinates.mat"),...
'imagesPixelCoordinates');
end
c4science · Help