can we extract optic disc from vessel image

good day all i have a extracted vessel image , i want to know whether i can extract optic disc from the same image. i dont want help with code..i just want to know how can i proceed ,i mean what method can be used. i looked in iris documentation but none of were suitable with me can somebody hel me below i have posted my image. i have got this by using fuzzy c means segmentation method..

[Dishant Arora]

how would you differentiate an optic disk??

[vidya]

from optic head nerve..or by using a 80*80 window.

[Ashish Uthama]

I guess what Dishant is trying to say is that for those of us not familiar with this domain, another image pointing out the region you want to extract (manually delineated) will help. I cannot be sure what part of the above image is the optic disk or the optic head nerve.

Answer 1

[Image Analyst]

This is what I'd try, assuming all you have to work with is the binary image.

1. do a hole fill
2. subtract the original binary image
3. filter out blobs that don't meet size and circularity and solidity criteria.

Of course it might be better if you started with the original color image and used some of the published methods in section 20.5 here: http://iris.usc.edu/Vision-Notes/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models

[vidya]

thanks a lot for ua reply sir i vl c how can i do with above method

[Image Analyst]

I gave you the algorithm. For 1, use imfill(). 2 is just regular subtraction. For 3, see my Image Segmentation tutorial http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862 where I filter blobs based on some criteria. It's only a few lines of code that I'd like to see if you can do it yourself. If you really can't figure it out, let me know and I'll help.

But like I said, why are you trying to find the optics disk from the binary image of the vessels, which is probably not best. The algorithm I outlined will only work for images like you posted where the optic disc is encircled by the vessel network. If it's not an enclosed circle, that won't work.

[vidya]

u are right sir..fortunately all my vessel images consits of a optic disc circle..extracting optic dis is also a part of my work so i thought i can somehow extract using the vessel image.

[Image Analyst]

Is that how it's done in all the published papers?

[vidya]

no sir.but most of the papers first extract vessels and then extract the optic disc...so now however i have extracted blood vessels i thought i can continue with the same

[Image Analyst]

You can continue with the optic disc extraction method outlined in the paper you are following. I don't know what it is but I think you do. Good luck.

[shakira idrees]

plz send me blood vessels extraction from retina matlab code

[Sidra Aleem]

@ Image Analyst I have done "retinal blood vessels segmentation". However, segmentation result is not so good. Below is attached the image having manual segmentation obtained from Drive database and segmentation of same imageusing my code.

I have done segmentation using following papers. http://manzaramesh.in/Publication/05-10-2013/Retinal%20blood%20vessels%20extraction%20using%202D%20median%20filter%20(2).pdf https://pdfs.semanticscholar.org/64c6/1b4e6214a6c6ab4e325794f41bae0109e6d9.pdf

I cannot attach my input image here as its in .tif format. I downloaded it from DRIVE DATABASE. Below is the link of DRIVE database.

http://www.isi.uu.nl/Research/Databases/DRIVE/

Below is attached my .m file. Please help me to improve segmentation. Kindly guide me to improve my results.