How to detect white regions in image

7 次查看（过去 30 天）
Pat
Pat 2014-11-11
评论： Pat 2014-12-7
I have a code below to detect exudates from retina Images
clc;
close all;
clear all;
workspace; % Display workspace panel.
% filename = 'C:\Documents and Settings\tk2013\My Documents\Temporary
% stuff\fundus.jpg';
rgbImage = imread('2.jpg');
[rows columns numberOfColorPlanes] = size(rgbImage);
subplot(3, 3, 1);
imshow(rgbImage, []);
title('Original color Image');
set(gcf, 'Position', get(0,'Screensize')); % Maximize figure.
tic;
redPlane = rgbImage(:, :, 1);
greenPlane = rgbImage(:, :, 2);
figure,imshow(redPlane)
K = imadjust(redPlane);
figure
imshow(K)
SE = strel('rectangle',[7 5]);
BW3 = imdilate(K,SE);
figure,imshow(BW3)
s=strel('square',12);
h=(imclose(BW3,s));
figure,imshow(h)
greenPlane=h;
[pixelCountsG GLs] = imhist(greenPlane);
% Ignore 0
pixelCountsG(1) = 0;
% Find where histogram falls to 10% of the peak, on the bright side.
tIndex = find(pixelCountsG >= .1*max(pixelCountsG), 1, 'last');
thresholdValue = GLs(tIndex)
binaryGreen = greenPlane>thresholdValue;
binaryImage = imfill(binaryGreen, 'holes');
% Get rid of blobs less than 5000 pixels.
binaryImage = bwareaopen(binaryImage, 5000);
figure,imshow(binaryGreen)
but the final output is only black,kindly help to extract the exudates from the above code,I have attached the images,I tried with different thresholds but could not get answer

请先登录，再回答此问题。

回答（2 个）

Image Analyst
Image Analyst 2014-11-11
So much wrong but I don't have time to fix it all. For starters, comment out the bwareaopen() function. And plot the histogram so you can see its shape.
  2 个评论
Pat
Pat 2014-11-12
Sir in the link u have given the code for green channel for fundus image http://www.mathworks.in/matlabcentral/newsreader/view_thread/273785
I tried it for red channel but could not process ,can u tell what must be changed in your code for processing red channel
Image Analyst
Image Analyst 2014-11-12
The red plane will have the least contrast. The blue or green channel will have more. That's why in my code I chose the green channel. But then you inserted code to adjust and do morphology on the red channel and then stick that modified red channel image into a variable deceptively named greenChannel.
I'm not going to have time to look at this for several days, if at all, so I ask you to look at algorithms published here http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models in Section 20.5, where people show how they've successfully done it. There is no need to invent your own algorithm when people have been working for months or years on algorithms and have published them for you to simply implement.

请先登录，再进行评论。

Siam
Siam 2014-11-12
编辑：Siam 2014-11-12
Your question is not clear. According to your code you need to concentrate (threshold and bwareaopen). You will get exactly what you are looking for.
  12 个评论
显示 10更早的评论
Siam
Siam 2014-11-17
Are you still having issues to get the result?

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Contrast Adjustment 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by