segmenting the lung region

i have converted the CXR to binary image, my aim is to obtain only those lung region,

i have used the same code as lung segmentation.m posted by @Image Analyst which was off great help, i would like further asisst in extracting only the lung region.

Any help is appreciated. Thankyou

[Simon Chan]

Did you try function imclearborder?

[ANUSHA H P]

l=imread('E:\project\dataset\CHNCXR_0042_0.png');
k=imresize(l,[256,256]);
p=imnoise(k,'salt & pepper',0.04);
h=medfilt2(p);
d=adapthisteq(h);
e=imbinarize(d);
f = imclearborder(e);
f = bwareafilt(e,5);
f = ~bwareaopen(~f, 1000);

i have used this code..

Answer 1

[Image Analyst]

Try

mask = imclearborder(mask); % Get rid of blobs touching the edge of the image.
mask = bwareafilt(mask, 2); % Take the largest 2 of the remaining blobs.

[ANUSHA H P]

Thankyou.

I have included this snippet in my code.

But when i give 2 for bwareafilt, i did not acquire the desired results.

l=imread('E:\project\dataset\CHNCXR_0042_0.png');
k=imresize(l,[256,256]);
p=imnoise(k,'salt & pepper',0.04);
h=medfilt2(p);
d=adapthisteq(h);
e=imbinarize(d);
f = imclearborder(e);
f = bwareafilt(e,2);
f = ~bwareaopen(~f, 1000);
N=edge(f,'canny');
figure,subplot(1,1,1),imshow(k),title('resized image');
figure,subplot(1,1,1),imshowpair(p,h,'montage');
figure,subplot(2,2,1),imshow(k),title('original image'),subplot(2,2,2),imhist(k);
subplot(2,2,3),imshow(h),title('filtered image'),subplot(2,2,4),imhist(h);
subplot(1,1,1),imshow(d),title('adaptive histogram equalization');
subplot(1,1,1),imshow(e),title('binarized');
subplot(1,1,1),imshow(f),title('removal of extra blobs');
subplot(1,1,1),imshow(BW),title('edge detected');

[Image Analyst]

It works. You were just doing a whole bunch of unnecessary stuff like adding noise, histogram equalization, edge detection, etc. Don't do that and just do what I say on your binary Image:

folder = pwd; % 'E:\project\dataset'
fullFileName = fullfile(folder, 'CHNCXR_0042_0.jpg');
rgbImage=imread(fullFileName);
[rows, columns, numberOfColorChannels] = size(rgbImage)
subplot(2, 2, 1);
imshow(rgbImage);
axis('on', 'image')
title('Original RGB image')
% Convert to gray scale.
if numberOfColorChannels == 3
    grayImage = rgb2gray(rgbImage);
else
    % It's already gray scale.
    grayImage = rgbImage; 
end
% Resize.
grayImage = imresize(grayImage,[256,256]);
subplot(2, 2, 2);
imshow(grayImage);
axis('on', 'image')
title('Resized gray scale image')
% Binarize
binaryImage = grayImage > 128;
subplot(2, 2, 3);
imshow(binaryImage);
binaryImage = imclearborder(binaryImage);
binaryImage = bwareafilt(binaryImage,2);
subplot(2, 2, 3);
imshow(binaryImage);
axis('on', 'image')
title('Lungs Only')

I don't even think the resizing is needed. Why are you doing that?

[ANUSHA H P]

As i am new to matlab, i was trying those commands sir.

Thankyou so much for the help, i tried the same code, it was off great help.

i tried the same code for a bunch of images, it did work for few , for rest it dint,

I have attached few input images for which i dint get the output, and a screenshot of output images in output folder.jpg. Can you please help me with the changes to be done.

[Image Analyst]

Yes, those are completely different images and will need a different algorithm. I suggest you look at published lung segmentation algorithms here:

• 20.7.2 Lungs, and Lung Cancer Image Analysis
• 20.7.2.1 Chest X-Ray Analysis
• 20.7.2.2 Lung Motion Analysis, Respiration, Breathing
• 20.7.2.3 Pneumonia, Lung Analysis, Flu, COVID
• 20.7.2.4 Emphysema, Lung Analysis
• 20.7.2.5 Airway Tree Structure
• 20.7.2.6 Bronchoscopy Systems, Bronchial Analysis
• 20.7.2.7 Thorax, Thoracic Analysis
• 20.7.2.8 Pulmonary Nodules, Lung Nodules
• 20.7.2.9 Tuberculosis Analysis, Tuberculosis Bacilli
• 20.7.2.10 Ribs, Chest X-Rays

[ANUSHA H P]

Thankyou sir for the reference.

Answer 2

[KSSV]

rgbImage = imread('https://in.mathworks.com/matlabcentral/answers/uploaded_files/858110/image.jpg') ;
% Removing the extra white background around the image (credit to Image
% Analyst)
grayImage = min(rgbImage, [], 3);
binaryImage = grayImage < 200;
binaryImage = bwareafilt(binaryImage, 1);
[rows, columns] = find(binaryImage);
row1 = min(rows);
row2 = max(rows);
col1 = min(columns);
col2 = max(columns);
% Crop
croppedImage = rgbImage(row1:row2, col1:col2, :);
BW = imbinarize(rgb2gray(croppedImage)) ;
[B,L] = bwboundaries(BW,'noholes');
imshow(label2rgb(L, @jet, [.5 .5 .5]))
hold on
for k = 1:length(B)
   boundary = B{k};
   plot(boundary(:,2), boundary(:,1), 'w', 'LineWidth', 2)
end

[ANUSHA H P]

hi, Thankyou for the response, this was of great help.

One more clarification,

I need the output to be in binary form i.e ROI.

I have attached a reference image named as ROI.png.