Threshold image after applying mask

Hi,

I have an image (object + background, uint16) and I remove the background of this image by using imbinarize. So the background has a value of zero and the object has its original values (1000 - 7000). Now I want to apply a threshold only on the object. I tried to change the zero values to nan, but that did not work. How can I only threshold the object, without all the zeros of the background?

Thanks!

Edit: I attached a photo of my file. Inside the red line is the object and outside is the background (black). I want to apply a threshold on the intensity of the object. I tried to remove the background by a mask, but when applying an automatic threshold, the background is still included. So hope someone can help.

Answer 1

[Image Analyst]

I think imbinarize() and graythresh() work on 1-D images. So if you want to determine the a threshold only considering the gray levels of the gray ring and not considering the black background, extract the ring pixels into a list and pass them to graythresh():

ringPixels = originalImage(originalImage > theThreshold); % No black pixels here
newThreshold = graythresh(ringPixels);

Since you've already masked the image by doing something like

maskedImage = originalImage; % Initialize
maskedImage(mask) = 0; % Erase background.

you could also do the first line like this

ringPixels = maskedImage(maskedImage > 0);

and then call graythresh();

[Ihiertje]

Thank you, this worked!

[Dua Arshad]

Sir, How can i apply this graythresh masking concept on the whole grayscale image. i have to segment the backgroud of fruits.

is it possible if i take first the binary image and make the background black and then use its mask to the gray image to remove background?

[Image Analyst]

You can do :

% Assume background is dark.  Threshold to find dark pixels.
backgroundMask = grayImage < 50; % or > or whatever value works for you instead of 50.
% Blacken background.
grayImage(backgroundMask) = 0; % Set background to 0.