how do i identify the background of binary image?

i have a binary image and I want to identify the background of the image? the image is bellow.

[naya taaj]

Does any one have any ideas??

[Image Analyst]

Not yet, probably because you posted the image as a .fig file which requires several extra steps for us to look at it. Why don't you simply post a screenshot as a .PNG (or other standard format) using the green and brown frame icon?

Basically though, the background in a binary image is the black pixels. They will have logical values of false, or numerical values of 0.

[naya taaj]

[naya taaj]

this is my image that i got from triangle thresholding method, i want to identify the background in it so i can remove it.

Answer 1

[Image Analyst]

In your case, the background is white. So you already have a background mask and you just need to tell us what "remove" means to you. Do you want to blacken it in the original image? If so, just do

% Mask the image using bsxfun() function to multiply the mask by each channel individually.
maskedRgbImage = bsxfun(@times, rgbImage, cast(~mask, 'like', rgbImage));

You can also use a gray scale image instead of rgbImage in the above code. Or, simply do

grayImage(mask) = 0;

[naya taaj]

@Image Analyst, i want to identify the objects in the image into background and foreground, then the resulting image(background identification result image) will be threshold to git rid of the background and obtain only the black objects. In my work i must identify and threshold this image and not the rgb or gray image. The resulting umage(after being identified then thresholded) will be white objects on black ground.

[Image Analyst]

So, once you have the mask image (that you have already posted), invert it, and assuming the blobs of interest are now white, you can simply call regionprops on it to get various measurements.

props = regionprops(mask)

[naya taaj]

@Image Analyst, ok, I understand what do you mean, but not all the objects considered as foreground, the completely black ones only, and the others small ones are considered as background, I hope you understand my explanation.

[naya taaj]

[naya taaj]

[naya taaj]

@Image Analyst, the 1st image is the background identification image, and the 2nd one is the background removal result.

[Image Analyst]

OK. Split apart those blobs with the code here: http://blogs.mathworks.com/steve/2013/11/19/watershed-transform-question-from-tech-support/ then call regionprops().

[naya taaj]

@Image Analyst, sorry for being bother you, but I must get this image before using this segmentation technique, because this image will be segmented by this technique after being processed as i mentioned, sorry again but i need your help.

[Image Analyst]

Use imread() to get the image before segmentation:

% "get this image before"
rgbImage = imread(filename);

Then do your processing:

% Now "being processed as i mentioned"
% Do whatever operations you do...
processedImage = ..... whatever......

Then finally do segmentation after processing.

% Segment the image by thresholding the processed image.
% "this image will be segmented by this technique after being processed"
binaryImage = processedImage < someThreshold;

[naya taaj]

@Image Analyst, seems like you does not pic my question, my question is how to implement 2nd image from 1st ?? If you can answer I'll be thankful to you.

[naya taaj]

Does anyone have any ideas??

[Image Analyst]

How did you get the second image? I assume you have some code to "implement" it, otherwise how could you have gotten it and posted it?

[naya taaj]

@Image Analyst,The 2 images i posted , i got them from the paper( screenshots).I want the code to implement them, once i got them from a code then my question is answered.

[Image Analyst]

You didn't post the original image, just one that was already thresholded (segmented). You said "I must get this image before using this segmentation technique." I can't help you get the original image. I'm sure you must have lots of original images to process, don't you? Just use one of your own.

And to separate the blobs, for that I gave you code (it's in Steve's blog).

I gave you code for

1. how to read in an image,
2. how to mask an image,
3. how to threshold (segment an image),
4. how to extract blobs based on certain attributes, and
5. how to split apart blobs.

So I really don't know what more you're asking for. Do you want me instead of you to implement Steve's watershed code? I really have no idea.

Please see my Image Segmentation Tutorial for a fully worked out segmentation example, as well as some other things like color segmentation, in my File Exchange: https://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862&sort=downloads_desc

[naya taaj]

@Image Analyst, thanks so much for your tries to help, i didnt post my original image because after many steps i got the image i posted( from triangle threshold) and no need for original image, i just asked about code to get the 2 images ( paper images) from the image i thresholded( by triangle) ,and about the watershed it is later step i am not reached yet.again thanks you try to understand and help me but as i said you cant reach what i meant.

[Image Analyst]

Did you see at the bottom of Steve's blog where it said "Get the MATLAB code"? If you click on that, it will run a javascript and all the code for the watershed is there. Of course, I don't know if it's exactly the process done in your paper so you might have to compare and use whichever one you want if they're different.

[naya taaj]

Can anyone tell me how to join mathwork newsgroup, or how i can communicate with a certain mathwork participant ???