How can I get the values of a circular region around a particular pixel?

How can I get the values of a circular region around a particular pixel?

Answer 1

[David Young]

Assuming:

• your image is grayscale and is called img;
• the central pixel is at column x and row y;
• the radius of the circle is r;
• you do not require weighting at the rim of the circle - that is, each pixel is either entirely inside or entirely outside the circle;

then you can do something like this

[xgrid, ygrid] = meshgrid(1:size(img,2), 1:size(img,1));
mask = ((xgrid-x).^2 + (ygrid-y).^2) <= r.^2;
values = img(mask);

where values will have a column vector containing the values in the circular region.

[Nafeesath Musfira]

Thanks a lot, hope this will work.

[Jennifer Bernier]

I'm doing something similar but with 5000 unique circular ROIs on a single image, and I can't figure out how to scale this method without using a loop (which takes a long time). Does anyone have any suggestions?

[Image Analyst]

5000 iterations of a for loop should take only a few microseconds Here is the timing on my system:

Elapsed time is 0.000010 seconds.

So, 10 microseconds. If yours is taking a lot longer, then there is something else that's taking the time, it's not the for loop itself.

Answer 2

[Image Analyst]

There's a FAQ on how to mask an image: http://matlab.wikia.com/wiki/FAQ#Image_Processing_Toolbox:

% Demo to have the user freehand draw an irregular shape over a gray scale image.
% Then it creates new images:
% (1) where the drawn region is all white inside the region and untouched outside the region,
% (2) where the drawn region is all black inside the region and untouched outside the region,
% (3) where the drawn region is untouched inside the region and all black outside the region.
% It also (4) calculates the mean intensity value of the image within that shape,
% (5) calculates the perimeter, centroid, and center of mass (weighted centroid), and
% (6) crops the drawn region to a new, smaller separate image.