MATLAB: adding columns to an empty matrix only when condition is true - matlab

Question

I have a image (orig) and a corresponding binary mask (maskD) composed of vertical streaks/columns of values zero or one. I am trying to make a third matrix (streakTemp) composed of only those columns in the image which have a 1 value in the mask. I'm using the code below and for some reason its giving me as an output both the columns I'm looking for and then zero values where the mask value is 0...so my output has the same x length as my input image...it should be shorter with the mask values of zero excluded. Not sure what I'm doing wrong..any thoughts? Thanks!
streakTemp=[];
for i=1:x
if maskD(1,i)==1
streakTemp(:,i)=orig(:,i);
end
end
imtool(streakTemp);

Answer 1

This variant of your code should work:
streakTemp=[];
j=1;
for i=1:x
if maskD(1,i)==1
streakTemp(:,j)=orig(:,i);
j=j+1;
end
end
The problem you have is that the i index always corresponds to the original matrix column, thus it won's skip the column even if the mask condition is not met.

Answer 2

I think this is what you need:
orig = reshape(1 : 20, 4, 5)
maskD = [1 0 0 1 1; 1 0 0 1 1; 1 0 0 1 1; 1 0 0 1 1]
mask1D = maskD(1, :)
x = 5;
streakTemp=[];
for i=1:x
if maskD(1,i)==1
streakTemp(:,i)=orig(:,i);
end
end
streakTemp
streakTemp2 = orig(:, logical(mask1D))
It outputs:
orig =
1 5 9 13 17
2 6 10 14 18
3 7 11 15 19
4 8 12 16 20
maskD =
1 0 0 1 1
1 0 0 1 1
1 0 0 1 1
1 0 0 1 1
mask1D =
1 0 0 1 1
streakTemp =
1 0 0 13 17
2 0 0 14 18
3 0 0 15 19
4 0 0 16 20
streakTemp2 =
1 13 17
2 14 18
3 15 19
4 16 20
Here is more on logical indexing.
Note that your maskD does not have to be a matrix at all; you only need to store the first line, which is why I use mask1D.

Answer 3

You shouldn't be using loops here. Straight up logical indexing is totally fine for your purposes:
streakTemp = orig(:, maskD(1,:) == 1);
Remember that maskD is a mask that's the same size as your original image, so we only need to access the first row to do the check. Simply put, this takes a look at all columns where maskD is equal to 1 then uses the corresponding locations to subsample from your orig matrix to create a new matrix that removes all columns that are not desired.