image to binary issues

Hi,

i want to run a single image through multiple thresholds. Initially, i had separate lines of code for each threshold (im2bw(img,.1), im2bw(img,.2) and so on). Now i want to run the image between thresholds ranging from 0:.001:.5. thats a total of 500 hundred values. How do I write that in few lines rather than write 500 lines?

I have tried for loop

thld = 0 : .001 : .5
for i=1:500
    a(i)=im2bw(img,thld)
end

but this doesn't work.

Thank you for the help

Answer 1

[Guillaume]

Nearly there. Use cell arrays for storing your images or a 3d matrix:

thld = 0 : .001 : .5;
for idx = 1:numel(thld)
  a{idx} = im2bw(img, thld(idx));
end

or

thld = 0 : .001 : .5;
a = zeros([size(img), numel(thld)]); %better preallocate huge array
for idx = 1:numel(thld)
  a(:,:, idx) = im2bw(img, thld(idx));
end

[jchris14]

This works well. But now need to find the difference in the number of black pixels to white pixels.

I used to do,

wpix=sum(a(:)); %a is the im2bw function

bpix=numel(a)-wpix;

this wont work with cell arrays unfortunately. Any ideas? I tried using cell2mat with no luck unless im using it wrong.

thanks again

[Image Analyst]

First of all, img better be a double image. If it's a uint8 image, or even a double image that was created from a uint8 image with something like imdouble(img) or img/max(img(:)) then you're not going to get more than 256 threshold, not 501. You can do this:

thld = 0 : .001 : .5;
numWhitePixels = zeros(1, length(thld));
numBlackPixels = zeros(1, length(thld));
for idx = 1:numel(thld)
  thisImage = im2bw(img, thld(idx));
  numWhitePixels(idx) = sum(thisImage(:));
  numBlackPixels(idx) = numel(thisImage) - numWhitePixels(idx);
  % Save to a cell array only if 
  % you need all images after the loop has exited.
  a{idx} = thisImage;
end

[jchris14]

I am trying to quantify a TIF image. But this is the error i get when i run the code.

Error using im2bw Expected input number 1, I, X or RGB, to be one of these types:

single, uint8, uint16, int16, logical, double

Instead its type was char.

Error in im2bw>parse_inputs (line 79)

validateattributes(varargin{1},...

Error in im2bw (line 38)

[A,map,level] = parse_inputs(varargin{:});

Error in R1 (line 20) thisImage = im2bw(img

[Image Analyst]

Your img is a character string. Did you pass the filename to im2bw() instead of the image array? It should look like this:

img = imread(fullFileName);
binaryImage = im2bw(img,..............

[jchris14]

My code says

a= uigetfile('*.tif');
thld = 0 : .001 : .5;
numWhitePixels = zeros(1, length(thld));
numBlackPixels = zeros(1, length(thld));
for idx = 1:numel(thld)
thisImage = im2bw(a, thld(idx));
numWhitePixels(idx) = sum(thisImage)
numBlackPixels(idx) = numel(thisImage) - numWhitePixels(idx)
% Save to a cell array only if 
% you need all images after the loop has exited.
%a{idx} = thisImage;
end

[Guillaume]

Use more descriptive name than a for your variable and you'll see where you go wrong.

You ask for a filename with uigetfile but you never load the image. In addition even if you'd loaded the image in a, you would have overwritten it in the loop since you're also using a for storing the binary image.

Your code should be something like:

imgfilename = uigetfile('*.tif');
sourceimage = imread(imgfilename);
...
binaryimage = im2bw(sourceimage, thld(idx));
binaryimages{idx} = binaryimage;

[Image Analyst]

Yep, like I said, the badly-named "a" is the culprit. It's a string just like I said. Did you use imread() like I suggested in my prior comment? No, not in that code you just posted. Why don't you use my suggestion? It will be alright then if you do.

[jchris14]

Thank you for the help. I will use them.