Problem:
I am trying to construct a vtk polydata model from a CT Nifity volume using marching cube method.
What I did:
So far I can produce a perfectly-to-scale skull model using vtk's polydata writer. However, the skull.vtk is rotated and translated rigidly when compared to the original ct.nii volume. I understand that Nifities have a QForm matrix to map voxel data to real world and vktPolyData do not have this data explicitly. However, the result of applying the QForm matrix to the vtkPolyData is not even close to perfect overlapping.
Does anyone know why this happens?
I think the reason is that the y-axis is inverted in VTK. Most likely applying a rotation matrix of [-1 0 0; 0 -1 0; 0 0 1] (row-major) will solve your problem.
A quick and easy check is to use an external tool, e.g. 3D Slicer or Paraview, to apply the rotation. For example in paraview, you can do the following:
-Load the polydata and your image in Paraview.
-Change the rotation to (180, 180, 0) degrees (equivalent to above rotation matrix)
-See if the polydata and image line up
I have attached an example image that shows how to do this in paraview. The red is the result of applying rotation to the original blue polydata. The location of the rotation parameters is given in the green box.
Related
Let's say i have an image like that one:
After some quick messing around, i got a binary image of the axe, like that:
What is the easiest/fastest way to get the contour of that image using GNU/Octave?
In Octave you can use bwboundaries (but I will welcome patches that implement bwtraceboundaries)
octave:1> pkg load image;
octave:2> bw = logical (imread ("http://i.stack.imgur.com/BoQPe.jpg"));
octave:3> boundaries = bwboundaries (bw);
octave:4> boundaries = cell2mat (boundaries);
octave:5> imshow (bw);
octave:6> hold on
octave:7> plot (boundaries(:,2), boundaries(:,1), '.g');
There are a couple of differences here from #Benoit_11 answer:
here we get the boundaries for all the objects in the image. bwboundaries will also accept coordinates as input argument to pick only a single object but I believe that work should be done by further processing your mask (may be due to the jpeg artifacts)
because we get boundaries for all objects, so you get a cell array with the coordinates. This is why we are using dots to plot the boundaries (the default is lines and it will be all over the image as it jumps from one object to other). Also, it is not documented whether the coordinates given are for the continuous boundary, so you should not assume it (again, why we plot dots).
the image that is read seems to have some artifacts, I will guess that is from saving in jpeg.
You can use bwtraceboundary in the Image package. Here is the Matlab implementation but that should be pretty similar using Octave:
First estimate starting pixel to look for boundary and then plot (BW is the image). (Check here )
dim = size(BW);
col = round(dim(2)/2)-90;
row = min(find(BW(:,col)));
boundary = bwtraceboundary(BW,[row, col],'N');
imshow(BW)
hold on;
plot(boundary(:,2),boundary(:,1),'g','LineWidth',3);
Output:
I'm working on a PIV-Workflow and I'm currently pre-processing the images. I need to get rid of the perspective distortion in the images. I do have the "image processing toolbox" and the "camera calibrator". I already got rid of the lens distortion with "undistortImage();" and the cameraParams object, which is inferred through a chessboard pattern.
First Question: Is it possible to use the cameraParams object to distort the image perspectively, so that my chessboard is rectified in the image?
Second Question: Since I were not able to use the cameraParams object, I tried to use the transformation functions manually. I tried to use pairs of control-points (with cpselection tool, the original image and a generated chessboard-image) and the fitgeotrans(movingPoints, fixedPoints, 'projective'); function to get my tform-object. However I always get the error message:
Error using fitgeotrans>findProjectiveTransform (line 189)
At least 4 non-collinear points needed to infer projective transform.
Error in fitgeotrans (line 102)
tform = findProjectiveTransform(movingPoints,fixedPoints);
I tried a lot of different pairs of control-points (4 pairs or more). But I'm still getting this error. I believe I must overlook something here.
Any help is appreciated, thank you.
Stephan
If you are using one of the calibration images, then all the information you need is in the cameraParams object.
Let's say you are using calibration image 1, and let's call it I.
First, undistort the image:
I = undistortImage(I, cameraParams);
Get the extrinsics (rotation and translation) for your image:
R = cameraParams.RotationMatrices(:,:,1);
t = cameraParams.TranslationVectors(1, :);
Then combine rotation and translation into one matrix:
R(3, :) = t;
Now compute the homography between the checkerboard and the image plane:
H = R * cameraParams.IntrinsicMatrix;
Transform the image using the inverse of the homography:
J = imwarp(I, projective2d(inv(H)));
imshow(J);
You should see a "bird's eye" view of the checkerboard. If you are not using one of the calibration images, then you can compute R and t using the extrinsics function.
Another way to do this is to use detectCheckerboardPoints and generateCheckerboardPoints, and then compute the homography using fitgeotform.
On a shape from a logical image, I am trying to extract the field of view from any point inside the shape on matlab :
I tried something involving to test each line going through the point but it is really really long.(I hope to do it for each points of the shape or at least each point of it's contour wich is quite a few times)
I think a faster method would be working iteratively by the expansion of a disk from the considered point but I am not sure how to do it.
How can I find this field of view in an efficient way?
Any ideas or solution would be appreciated, thanks.
Here is a possible approach (the principle behind the function I wrote, available on Matlab Central):
I created this test image and an arbitrary point of view:
testscene=zeros(500);
testscene(80:120,80:120)=1;
testscene(200:250,400:450)=1;
testscene(380:450,200:270)=1;
viewpoint=[250, 300];
imsize=size(testscene); % checks the size of the image
It looks like this (the circle marks the view point I chose):
The next line computes the longest distance to the edge of the image from the viewpoint:
maxdist=max([norm(viewpoint), norm(viewpoint-[1 imsize(2)]), norm(viewpoint-[imsize(1) 1]), norm(viewpoint-imsize)]);
angles=1:360; % use smaller increment to increase resolution
Then generate a set of points uniformly distributed around the viewpoint.:
endpoints=bsxfun(#plus, maxdist*[cosd(angles)' sind(angles)'], viewpoint);
for k=1:numel(angles)
[CX,CY,C] = improfile(testscene,[viewpoint(1), endpoints(k,1)],[viewpoint(2), endpoints(k,2)]);
idx=find(C);
intersec(k,:)=[CX(idx(1)), CY(idx(1))];
end
What this does is drawing lines from the view point to each directions specified in the array angles and look for the position of the intersection with an obstacle or the edge of the image.
This should help visualizing the process:
Finally, let's use the built-in roipoly function to create a binary mask from a set of coordinates:
FieldofView = roipoly(testscene,intersec(:,1),intersec(:,2));
Here is how it looks like (obstacles in white, visible field in gray, viewpoint in red):
I am trying to map my image point by point to 3 dimensional space.
For example, if my original image has intensity of 100 at location X, I want to plot this point in 3D location Y with intensity of 100. I want to repeat this steps for every point/pixel, and get a final image. My biggest problem is that I want to do it point by point.
I appreciate any comments or advice. Thank you.
=======================
p.s.
As I was writing this question, I just came up with an idea. I know how to print 'whole' image into certain location/shape in 3D by using warp() function. Instead of using my whole image as an argument to warp function, if I give one point intensity value and one 3D point as arguments for warp function, and repeat this steps for every image point, will I get a descent looking final image in 3D? If there is a better function to use, please let me know.
Sounds like you are looking for scatter3:
I = imread('cameraman.tif');
[x y]=meshgrid(1:size(I,1), 1:size(I,2));
scatter3(x(:),y(:),I(:),15,I(:),'filled');
axis tight; colormap gray
And this is what you get (after some changes to view point):
PS,
I used a single scatter3 command to plot all the points at once. You may (I have no idea why you would like to do so) do it one by one
figure;
for ii=1:numel(x)
scatter( x(ii), y(ii), I(ii), 15, I(ii), 'filled');
hold on; % need this!
end
axis tight; colormap gray;
I am trying to generate the following "effect" from a basic shape in MATLAB:
But I don't even know how this process is called. Let's say I have an image containing the brown shape, what I want is generate the contours outside of it, that get smoother as they get bigger.
Is there either a name for this effect, a function to do this in MATLAB or an algorithm that does it from scratch?
thanks
I think you are looking for bwdist.
The image you are displaying looks like the positive part of a distance function from the boundary of your shape. You can perform this easily in Matlab using the examples on the aforementioned manual page.
Try this:
I = imread('brown_image.png');
I_bw = (rgb2gray(I) > 0); % or whatever, just so I_bw is 1 in the 'brown' region
r = 10;
se1 = strel('disk', r);
se2 = strel('disk', r-1);
imshow(imdilate(I_bw, se1) - imdilate(I_bw, se2))
Requires image processing toolbox, but the basic idea is to dilate the image twice with dilation elements that differ by 1 (or however thick you want the contours to be) and subtract the result of the smaller one from the bigger one. You could then color them however you want.