Detecting thin edges of the connected cell

Hello Everyone! I have a B/W image of a cell and want to extract a thin edges out of it. I tried imclose() but it gives the closed object. Please find the source image, target image (Image which I want) and image with imclose() operation. Yellow line in the Targetimage.jpg is the output, which I want. Thanking you in anticipation.

Answer 1

[Image Analyst]

First call imfill() on your source image to get rid of internal black spaces:

filledImage = imfill(sourceImage, 'holes');

Then separate the blobs with watershed. For that see Steve's blog. Attach your code if you have any problems with Steve's algorithm.

[Mahi]

Hi Image Analyst, Thanks very much for your prompt response. I am new in this Image Processing field and learned a lot of things from your blogs and answers. Thank you very much once again. I implemented algorithm and it worked correctly as well. However, I have an extended question: I have got a braking line of that mask with more than one pixel thick. I want a thin breaking line, one pixel thick. Please find the attached output image of mine as well. Is there any technique, by which I can do this- one pixel thick boundary line

[Image Analyst]

You can deal with them one blob at a time. Dilate the one that you want to expand to cover the separation gap with imdilate().

[Mahi]

Hi Image Analyst, The method which you have suggested, worked perfectly. However, the segmentation was not the perfect, as the segmented line was the straight line. The main edge information was lost. Edge information file is attached with main_image.png in this question. Kindly guide me in this regard. Thanks.

[Image Analyst]

What's wrong with a straight line? If you don't like it, use imline() or imfreehand() to manually trace out a break line.