How can I find the overlap between two binary volumes?

I'm trying to compare two volumes consisting of 3D white shapes on a black background using matlab 2013 for windows.

1) I want to know how much they overlap over all. Correct me if I'm wrong but I think I can use the nnz function to do that.

2) I want to know where the volumes do not overlap. I have no clue how to accomplish this.

Any help is appreciated.

Answer 1

[Image Analyst]

Define "find the overlap". To get the overlap image:

overlapImage = binaryVolume1 & binaryVolume2;
% To count number of pixels
numOverlapPixels = nnz(overlapImage);

[Dalas]

Given two 3d arrays with equal number of x,y,z coordinates (256 by 256 by 256 for example) where each coordinate can either be a one (white) or a zero (black). I want to find out how many coordinates match between the two arrays and make a new binary image stack or volume which shows where the original two volumes have the same values and where they have non-matching values.

[Image Analyst]

% Create the volumes:
binaryVolume1 = false(256,256,256);
binaryVolume2 = false(256,256,256);
% "Light up" pixels
for k = 1 : size(x)
  row = y(k);
  column = x(k);
  slice = z(k);
  binaryVolume1(row, column, slice) = true;
end

Then do for the other volume, binaryVolume2, with its x,y, and z. Then use the code above in my answer to find overlap or matching voxels.

[Dalas]

I dont need to create the volumes I already have them i just need to analyze them.

I'm looking at mri brain scans by the way just so you have some context of what im working on.

I geuss the volumes arn't actually volumes either their simply stacked 2d "brain slices". We arnt using 3d mr image aquisition. But I think for matlabs purposes its the same as a 3d volume.

Would xor work on 3d arrays?

[Dalas]

We're ignoring tissue differences and converting the images into a binary image to focus on brain volume at the moment rather than pathology.