how comparing pixelwise a patch of pixel with distinct intensity on image with Euclidean distance?

I need to create two pixel patches of size 5×5 with constant intensity values μ=150 for X location and η ∈[0,255] for Y location, and then comparing this pixel patch pixelwise with image by the squared l2-distance(Euclidean distance). Comparison of the pixel patches over the one image should perform 10,000-times for every value of η∈[0,255]to fortify this experiment with sufficient statistics.

Then I need to create a plot of this experiment that shows the average distance of the two pixel patches and the standard deviation.

Any help how can I do it in matlab?

The desired result is like attached file.

[Image Analyst]

Something is missing in your explanation. First of all, where is the image? Next, you're comparing a 5x5 window of the image with a 5x5 window with constant value μ=150, with the comparison being (I guess) sum((pixelValue-150)^2) where the sum is over the 25 pixels in the window, and I presume the window slides along the image. So, where does η come into this computation??? I have no idea. Next, you say you need to do this 10,000 times? WHY??? You didn't say the image changed, nor did you say that mu changes, so what changes from one time to the next? Maybe η but like I said, that's not involved in the computation so it won't affect the results. So every one of the 10,000 computations will give you the same image and the same sum when averaged over all possible window locations.

[Neda]

The image is dicom(ultrasound img). In fact, I am following an experiment on the paper attached link on page 5. Maybe my understanding is wrong. The pixel patch on (x,y) location with intensity(150,[0,255]) should scan over the image. paper

[Image Analyst]

It doesn't look like I could code up that paper for you in 5 minutes, so I suggest you ask the author for code.