Extract contour of an image and plot it

Hi,

I want to extract the contour from an image and plot the contours assuming the center of the image as the (0,0) coordinate system.

I have attached my image too.

Thanks for your help.

Answer 1

[Walter Roberson]

im = rgb2gray( imread('Picture4.png') );
contour(im,4)

There is an obvious difference between specifying 2 (contour levels) or 3. However, you have to be looking fairly closely to tell the difference between 3 vs 4. You probably will not be able to tell the difference between 4 and 5.

The image appears to be consistent with the possibility that someone converted an image to black and white, and then write the image as jpeg, and then read in the jpeg and wrote it as png. That is, the values that are not 0 (black) or 255 (white) occur only in boundary positions as-if they only existed due to the way DCT handles sharp edges (it blurs them, like anti-aliasing.)

[Image Analyst]

I wonder if he's confusing intensity contours of a gray scale image with outlines/boundaries of a binary image?

[Easir Papon]

Thanks for your reply. And I am sorry I couldn't clarify the question properly. I just need to capture the countour of the boundary.

I know the maximum thickness of the shape is 0.4 mm. How can I convert each points of the contour (boundaries) into my scale and then I would plot it.

Thanks again.

[Walter Roberson]

The original source appears to have been a binary image. Is it possible to get that binary image directly instead of having been written as jpeg and then converted to png ? jpeg lossy compression is a Bad Thing when processing scientific images. You don't want to end up fitting features that turn out to be due to the way that JPEG's DCT smears sharp edges.

[Walter Roberson]

bw = imbinarize(im, 'ForegroundPolarity', 'dark');

Then you can use one of the two boundary tracing routines; see https://www.mathworks.com/help/images/boundary-tracing-in-images.html

Note that the boundaries get somewhat messy due to that thin white line. Everywhere that the thin white line happens to meet the outside, the boundaries have to be carefully traced around where the white line goes inside the black area. You might want to consider doing a imclose()

[Easir Papon]

I have attached the original image. I need to capture only the boundary.

[Walter Roberson]

Any chance you could use a lossless image format for the primary images?

[Easir Papon]

Unfortunately the last attached image is the only one I have.

[Image Analyst]

If all you have is one image, not hundreds, then why not simply trace it freehand, like with the atttached demo?

[Jean-Parfait]

I've tried to extract an image with the code you've suggested but it gives an invested response. Please can you suggest a code that will give a non-reversed extracted image ?

[Jean-Parfait]

Thank you

[Image Analyst]

@Jean-Parfait My demo does not reverse/invert the image intensities. I think you must have changed something. Nonetheless, I've updated a few lines in the demo and the updated code is attached.