MATLAB Answers

How to remove candidates which are present on blood vessels?

2 views (last 30 days)
Hello,
I am trying to remove pixels which are present on the blood vessels.I have two images ,one is the candidates image and the blood vessels image.I have studied that candidates won't occur on blood vessels.So, i have to eliminate the candidates which are present on blood vessels and get the output image.Can anyone help me with this? seg40.jpg
This is the bood vessel image and the candidate image is below as
output.jpg
I want to remove white dots on candidate image which are present on blood vessels in blood vessels image.
Thank you

Accepted Answer

KALYAN ACHARJYA
KALYAN ACHARJYA on 18 Nov 2019
Edited: KALYAN ACHARJYA on 18 Nov 2019
"I have two images ,one is the candidates image and the blood vessels image.I have studied that candidates won't occur on blood vessels.So, i have to eliminate the candidates which are present on blood vessels and get the output image"
Use appropriate logical Operations, for example
result=bloodvessels_image & ~candidate_image

  9 Comments

Show 6 older comments
Krishna Chaitanya
Krishna Chaitanya on 18 Nov 2019
That didn't help actually.But,I have tried this and it worked
del_region = bloodvessel_img & Candidate_img;
[r c] = find(del_region==1);
Candidate_img(r,c) = 0;
Thank you for your help.Can you help me in applying region growing to the first image,i.e, the blood vessel image?I want to fill the gaps within the blood vessels in first image.Thank you.
KALYAN ACHARJYA
KALYAN ACHARJYA on 18 Nov 2019
First one is much simpler solution. Anyway the problem is resoved with second answer.
You already apply the segmentation on the image, then why region growing?
Is it for vessels extraction for CAD diabetic retinopathy monitoring system?
Krishna Chaitanya
Krishna Chaitanya on 18 Jan 2020
The size and shape of detected MA candidates do not represent the true Microaneurysms hence,region growing should be used.
Yes,It is for CAD diabetic retinopathy monitoring system.

Sign in to comment.

More Answers (0)


Translated by