Blood Vessels recognition in retinal image

@grigo-Ym20du • Oct 22, 2024

Oct 22, 2024

1.1K

Hi, i'm new in this forum and i'd ask you some help; i'm trying to solve a problem in my university thesis but i need some ideas; i'm working with retinal images and i'm trying to detect the vessels in a automatic way with Matlab software; i tried to improve the image working in the contrast and brightness but now i have some problems about the accurate recognition of the single vessels. I tried to use K-means and fuzzy c-means algorithm but the results aren't so good.
Someone can give me help please??

Like 0 Replies 4

Replies

Welcome, guest

Join CrazyEngineers to reply, ask questions, and participate in conversations.

CrazyEngineers powered by Jatra Community Platform

Ramani Aswath

@ramani-VR4O43 • Sep 30, 2011

Have methods like the application of contrast colour optical filters to enhance contrast been tried?
0
grigo

@grigo-Ym20du • Oct 2, 2011

I used this code:
imm_temp=double(f_analyzed); m=min(imm_temp😀));
M=max(imm_temp😀));
betcontr_imm=255*(imm_temp-m)/(M-m);
betcontr_imm=uint8(betcontr_imm);
betcontr_immlog=mat2gray(log(1 + double(f_analyzed)));

I used two type of way to improve the contrast but the better solution is the logarithmic one.
0
grigo

@grigo-Ym20du • Oct 2, 2011

the smile is colon...
0
grigo

@grigo-Ym20du • Oct 2, 2011

the situation is this; i have a retinal image of the blood vessels and the seg structure with all the crossover and the bifurcations. I have to create an automatic algorithm that analyzes all the cross.. and bifurc... that improves the result of the tracking algorithm that gave me the seg structure with all the coordinates of the bifurc and crossover. I read a lot of papers but i don't understand how to proceed and what type of algorithm is best for my work.
To solve my problem i take all the single big and cross and analyze them alone, so i improve the contrast and use the Laplacian and logarithmic filters with this windows
w4=fspecial('laplacian',0);
w8=[1 1 1; 1 -8 1; 1 1 1];
wlog=fspecial('log');
The best solution i can see is the w8 but it depends in particular by the image that i'm analyzing.
0