How to remove the found image from the original image?

Hi, I am a beginner to MATLAB image processing.

I have an original fundus image as shown. And i have found the vessel tree as attached which I need to remove from the fundus image.

My question is how can I remove the vessel tree from the original image?

Thank you.

Answer 1

[Image Analyst]

You can invert the mask to get everything except the vessels, then mask:

% Get fundus mask, which is the opposite of the vessel mask
mask = ~vesselMask;
% Mask image by multiplying each channel by the mask.
maskedRgbImage = rgbImage .* cast(mask, 'like', rgbImage); % R2016b or later. Works for gray scale as well as RGB Color images.
% For R2016a or earlier, mask the image using bsxfun() function to multiply the mask by each channel individually.  Works for gray scale as well as RGB Color images.
% maskedRgbImage = bsxfun(@times, rgbImage, cast(mask, 'like', rgbImage)); % R2016a or earlier.

[ZWY]

Hi, i have tried the code suggested and get the image as attached.

Is there another way to remove the vessel and at the same time, maintaing the structure of the optic disk?

[Image Analyst]

The data from the optic disc basically includes the vessels if you wan tto think of it that way. Both blood vessels and nerves head toward the optic disc where they get bundled up and leave the eyeball out the back. Perhaps there might be some small nerve fibers going underneath the blood vessels. And it's possible that as they are all bundled into a rope-like structure, the vessels and nerves twist and wind around each other. So I'm not really sure what you want. I even did half my Ph.D. in Ophthalmic imaging with a laser scanning fundusscope.

If you want you can find all the bright blobs and then call bwconvhull

See how others define and find the optic disc here:

http://www.visionbib.com/bibliography/contentsmedical.html#Medical%20Applications,%20CAT,%20MRI,%20Ultrasound,%20Heart%20Models,%20Brain%20Models

[ZWY]

Hi, i have found one method which is

"The retinal vessel region segmented was transformed into NaN elements. The NaN elements were substituted with the new pixel values calculated using the four nearest non-vessel pixel values."

Assume the vessel tree in black and white image is bw. My question is how can I substitute the bw image with a grayscale index?

Here is my code:

bw=nan;

isnan(bw)

bw(isnan(bw))=impixel(imG, 2800, 1400) %imG is a green channel of an image

figure,imshow(bw);

Error: Unable to perform assignment because the left and right sides have a different number of elements.

[Image Analyst]

If you want to "fill in" the vessels with the weighted average of the nearby non-vessel gray levels, you can use regionfill.

[ZWY]

Thank you