If I have :
for i=1:n
for j=1:m
if outputImg(i,j) < thresholdLow
outputImg(i,j) = 0;
elseif outputImg(i,j)> thresholdHigh
outputImg(i,j) = 1;
end
end
end
or even worse :
for i=1:n
for j=1:m
for k=1:q
% do something
end
end
end
How can I achieve this differently , without for ?
Instead of the first loop you can use logical conditions, such as:
outputImg(outputImg<thresholdLow)=0;
outputImg(outputImg>thresholdHigh)=1;
There are of course many other equivalent ways to get that using logical operators...
For the second loop you need to be more specific, but I think you got the grips of the logical conditions trick.
For a general solution, look into ndgrid which in your second case you could use like this:
[i j k] = ndgrid(1:n, 1:m, 1:q);
ijk = [i(:) j(:) k(:)];
Then you can traverse the list of combinations of i, j, and k, i.e. now ijk to parameterize your thresholding statements.
ijk. You can simply use one for loop to go through the length of i which is the same as the length of k and of j to index these vectors. This is perhaps what @Leonid Beschastny meant by vectorizing your code.
[a1 a2 ... aN] = ndgrid(1:p); for variable N's. How to do this?If you use binary matrix:
index_matrix = (outputImg < thresholdLow);
The following hold:
index_matrix(i,j) == 0 iff outputImg(i,j) < thresholdLow
index_matrix(i,j) == 1 iff outputImg(i,j) > thresholdLow
for the second loop you can always use matirx over for loop
