after we obtain a binary image ...is there any way to strong the edge??

hi after we obtain a binary image with edge for example with otsu method ...is there any way to strengthen these edges??

[Matt J]

@sara, When you post questions, the "Tags" field is not meant to contain a duplicate of your question. It is meant to contain useful keywords that people can use for search purposes, e.g., 'edge', 'otsu', etc...

[sara]

ok Matt

thanks

[Image Analyst]

I don't know what that means - to "strong" an edge. Strong is not even a verb. Please get another English speaker to translate.

[Salaheddin Hosseinzadeh]

Hi Image Analyst,

I believe she want's to stress/emphasize the edges. I'm afraid neither me nor Sara are native English speaker, but in our native language that's what strong means!

Would you please tell me how to count the coins below?! I thought by detecting their edges (lower right) it's going to be easy. I wrote something to eleminate the neighboring white pixels except one, so 1 white pixel of each coin remain, but the closest I could get was 9 coins, while it's only 4!

Regards,

SH

[Image Analyst]

Salaheddin: If she wants the "strength" of the edges, she can use imgradient() instead of edge() and just look at the value of the gradient image. If she wants to "enhance" edges to make them more contrasty, then she can use conv2() or imfilter() with a kernel like (-1,-1,-1;-1,17,-1;-1,-1,-1]. That's basically a Laplacian (edges) plus the original image. So it adds an edge (high pass) image, like imgradient(), to the original image, making the edges more prominent.

For the coin question, you should not use edge(), you can use thresholding. See my Image Segmentation Tutuorial - it uses coins. http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862

[Salaheddin Hosseinzadeh]

Hello Image Analyst.

Thanks a lot for the quick reply and the link to your demos.

I hope I live long enough, to learn all these cool things and even more :)

Many Regards,

SH

[Iain]

strengthen = to make stronger

strong, used like sara used it could equally have been a typo for storing as well as "make stronger"

[sara]

thanks lain

Answer 1

[Image Analyst]

To make edges more contrasty, use a high boost filter:

kernel = [-1,-1,-1;-1,17,-1;-1,-1,-1]/9;
highBoostFilteredImage = imfilter(double(grayImage), kernel);
imshow(highBoostFilteredImage, []);

That's one way, anyway. There are others.

[Image Analyst]

After reading your "Answer" I really don't know what you want to do. Perhaps you can find a paper here that did the same thing you would like to do: http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models

[sara]

thanks Image Analyst after obtaining enough information I will come back soon and ask my question in a clear way.

Answer 2

[sara]

thanks Image Analyst and Salaheddin Hosseinzadeh

dear Image Analyst with your guidance and some modification that I did , I obtain a better result..but if you see there is a lesion around lung .I thouth if I use a filter that strong the edge maybe the result is better ...I used your seggustion

if true
 kernel = [-1,-1,-1;-1,17,-1;-1,-1,-1]/9;
highBoostFilteredImage = imfilter(double(grayImage), kernel);
imshow(highBoostFilteredImage, []); 
end

but I think it is not a good idea...Image Analyst and Salaheddin Hosseinzadeh do you have a good idea?

[Image Analyst]

I believe the discussion is taking place now in your duplicate thread.