It seems that matlab is rounding my negative values to 0 when stored in a matrix. How can I undo that? And is the cosine values in the right order?

read = dicomread('2.dcm')
[X, map] = dicomread('2.dcm');
%montage(X, map, 'Size', [2 5]);
info = dicominfo('2.dcm');
Y = dicomread(info);
%figure, imshow(Y);
%imcontrast;

Rows = info.Rows
Columns = info.Columns
Pixel_Spacing = info.PixelSpacing %in mm/pixel
Pixel_Spacing_x = info.PixelSpacing(1)
Pixel_Spacing_y = info.PixelSpacing(2)
Patient_Position = info.ImagePositionPatient
Sx = double(info.ImagePositionPatient(1))
Sy = double(info.ImagePositionPatient(2))
Sz = double(info.ImagePositionPatient(3))
Image_Orientation = info.ImageOrientationPatient
F12 = Image_Orientation(1)
F22 = Image_Orientation(2)
F32 = Image_Orientation(3)
F11 = Image_Orientation(4)
F21 = Image_Orientation(5)
F31 = Image_Orientation(6)
Delta_r = Rows*Pixel_Spacing_x %in mm
Delta_c = Columns*Pixel_Spacing_y %in mm
%r: Row index to the image plane. The first row index is zero.
%c: Column index to the image plane. The first column is index zero.
r = 0.0
c = 0.0

A = double([(F11*Delta_r) (F12*Delta_c) 0 Sx; F21*Delta_r F22*Delta_c 0 Sy; F31*Delta_r F32*Delta_c 0 Sz;0 0 0 1])
B = double([r;c;0.0;1.0])
P = double(A)*double(B)

>> Sy

Sy =

   -27.721523479739002

>> A

A =

       0     0     0   175
       0   222     0     0
       0     0     0   192
       0     0     0     1

The value of Sy should be in the second row the fourth column.

What can I do to stop the rounding to 0? also, despite having double, in the matrix there is always rounding to nearest integer. I changed the formats but the problem is still there.

Answer 1

[Dehua Kang]

Hi ,Hoda Hatoum I think you may make a mistake. I think that F12 = Image_Orientation(4); F22 = Image_Orientation(5); F32 = Image_Orientation(6); F11 = Image_Orientation(1); F21 = Image_Orientation(2); F31 = Image_Orientation(3); cause the vector of ImageOrientationPatient is [1 0 0 0 1 0],so the A should be 500 0 0 -250 0 500 0 -140 0 0 0 -20 0 0 0 1 and P is -250 -140 -20 1

[HH0825]

Thank you for pointing that out!

Answer 2

[Walter Roberson]

You used double() after you construct A. Some component that you are putting into A is one of the unsigned integer data types, and MATLAB is seeing that and converting everything around it to the same unsigned integer data type before double() is called.

For example, hypothetically info.ImageOrientationPatient could be unsigned integer, which would make all of your F* variables unsigned integer, and then the expression building the array would all be converted to unsigned integer before double() was called.

[HH0825]

Hey Walter thank you for your answer. I only added double because it didn't work without it. I just tried to take all the doubles away except for the ones in the matrix multiplication line (P = double(A)*double(B)) and still I can't have the values right.

[Walter Roberson]

read = dicomread('2.dcm')
[X, map] = dicomread('2.dcm');
%montage(X, map, 'Size', [2 5]);
info = dicominfo('2.dcm');
Y = dicomread(info);
%figure, imshow(Y);
%imcontrast;

Rows = double(info.Rows);
Columns = double(info.Columns);
Pixel_Spacing = double(info.PixelSpacing); %in mm/pixel
Pixel_Spacing_x = info.PixelSpacing(1)
Pixel_Spacing_y = info.PixelSpacing(2)
Patient_Position = double(info.ImagePositionPatient);
Sx = info.ImagePositionPatient(1);
Sy = info.ImagePositionPatient(2);
Sz = info.ImagePositionPatient(3);
Image_Orientation = double(info.ImageOrientationPatient);
F12 = Image_Orientation(1);
F22 = Image_Orientation(2);
F32 = Image_Orientation(3);
F11 = Image_Orientation(4);
F21 = Image_Orientation(5);
F31 = Image_Orientation(6);
Delta_r = Rows*Pixel_Spacing_x; %in mm
Delta_c = Columns*Pixel_Spacing_y; %in mm
%r: Row index to the image plane. The first row index is zero.
%c: Column index to the image plane. The first column is index zero.
r = 0.0;
c = 0.0;

A = [(F11*Delta_r) (F12*Delta_c) 0 Sx; F21*Delta_r F22*Delta_c 0 Sy; F31*Delta_r F32*Delta_c 0 Sz;0 0 0 1]
B = [r;c;0.0;1.0]
P = A*B

[HH0825]

Thanks!

Answer 3

[Dehua Kang]

And here :Delta_r = Rows*Pixel_Spacing_x; %in mm Delta_c = Columns*Pixel_Spacing_y; %in mm

It may be wrong ,it should be :Delta_r=Pixel_Spacing_x;Delta_c = Pixel_Spacing_y; It is none of the Columns and Rows business.

Answer 4

[Dehua Kang]

This is my value of p :P =

   42.9688
  152.9688
  -20.0000
    1.0000

is not integer, it is double.

[Steven Lord]

1) Please stop posting comments that offer additional information or clarification as answers to the question. Post them as comments using the "Comment on this Question" or "Comment on this Answer" links under the main question or the answer to which you want to respond, respectively.

2) Set a breakpoint on the line where you define the variable A then execute your code. When you reach the breakpoint, look at the classes of the variables you use on that line to define A. The whos function will give you that information. If any of those variables are of an integer type (int8, int16, uint32, uint64, etc.) then the combined matrix will be of that integer type.

You can convert that combined matrix into double all you want after that point, but the data has already been lost in the conversion process. Once you crack an egg into a cake batter, remembering that you wanted to hard boil it for breakfast instead won't put it back in its shell.